Environmental WebAPIs State of Art

See discussions, stats, and author profiles for this publication at: https://www.researchgate.
net/publication/315712161
Web APIs for environmental data - state of the art investigation
Technical Report · June 2016

DOI: 10.13140/RG.2.2.17817.83047
CITATIONS READS
0 1,278
1 author:
Peter Taylor
The Commonwealth Scientific and Industrial Research Organisation
24 PUBLICATIONS 92 CITATIONS
SEE PROFILE
All content following this page was uploaded by Peter Taylor on 31 March 2017.
The user has requested enhancement of the downloaded file.

Web APIs for environmental data
State of the art investigation
Peter Taylor
30 June 2016
Web APIs for environmental data | 1

Contents
Acknowledgments ................................................................................................................................ 4
Executive summary .............................................................................................................................. 5
1 Introduction ................................................................................................................................... 7
1.1 Scope............................................................................................................................... 8
2 The need for Web APIs .................................................................................................................. 9
2.1 Why provide APIs? ........................................................................................................ 10
2.2 Who are the users of Environmental Web APIs? ......................................................... 12
2.3 Supporting new data-driven businesses ...................................................................... 13
2.4 What are the challenges for environmental data Web APIs? ...................................... 15
3 The nature of environmental data ............................................................................................... 16
4 Related domain APIs .................................................................................................................... 18
4.1 National Oceanic and Atmospheric Administration (NOAA) ........................................ 18
4.2 UK Met Office – DataPoint ........................................................................................... 24
4.3 Forecast.io .................................................................................................................... 29
4.4 USGS Water Data Services ............................................................................................ 30
4.5 Atlas of Living Australia ................................................................................................ 32
4.6 OGC SensorThings API .................................................................................................. 37
4.7 UK National Biodiversity Network ................................................................................ 39
4.8 GeoServices API ............................................................................................................ 39
4.9 Global Change Information System .............................................................................. 41
4.10 UK Environmental Agency – Bathing Water Quality API ............................................ 43
4.11 CKAN ........................................................................................................................... 44
4.12 Data Access Protocol (DAP), OpenDAP, Hyrax ........................................................... 45
4.13 THREDDS ..................................................................................................................... 46
4.14 ERDDAP ....................................................................................................................... 47
5 Web API practices ........................................................................................................................ 48
5.1 Describing service interfaces ........................................................................................ 48
5.2 Specifications for API responses ................................................................................... 50
5.3 API management .......................................................................................................... 55
6 Conclusions .................................................................................................................................. 59
6.1 General ......................................................................................................................... 59

6.2 Technical ....................................................................................................................... 62
6.3 Modelling the domain .................................................................................................. 64
Appendix A 65
Summary of example data products from the Australian Bureau of Meteorology .............. 65
Summary of Web APIs ........................................................................................................... 72
References 74

Acknowledgments
This work has, for the most part, been funded through a Water Information Research and
Development Alliance (WIRADA) between CSIRO and the Australian Bureau of Meteorology. The
author wishes to thank Simon Cox (CSIRO), Dominic Lowe (Bureau of Meteorology) and Robert
Power (CSIRO) for their thorough reviews.

Executive summary
Environmental data provides huge value to society. It is used in many ways, from local planning
and monitoring, to global modelling and analysis. Organisations publish environmental data using
a range of techniques, but increasingly it involves the use of the Internet (web sites, email, social
media, mobile devices, automated notifications etc.). Publishing data for use in current and future
software is an important way value is gained from environmental data. Globally, organisations are
increasingly using Web APIs (Application Programming Interfaces) to supply data to multiple
audiences, and creating new businesses in the process.
This report summarises existing practices for the publication of environmental data using Web
APIs. The aim is to provide organisations with a well-informed basis to design consistent, cross-
organisation Web APIs to access a range of environmental data products. While this report was
conducted in a hydrology-specific activity (WIRADA), it takes a broader look at environmental data
to contribute to a cross-domain view of Web APIs.
The Web APIs covered in this report include a selection from the USGS, NOAA, the UK Met Office,
and the UK environmental agency, among others. There are also some commercial and open
source offerings that have no direct affiliation with environmental agencies. In our selection we
focussed on organisations with similar data holdings to the Australian Bureau of Meteorology and
those that had progressed to public facing Web APIs.
Generally the practices varied quite a lot across these organisations. The specifics of their Web API
depend on the time they were developed and their target domain. Some use older Web API
practices (e.g. simple key-value pair APIs), while others use newer approaches (e.g. linked data).
This raises an important point: organisations must find a balance between adopting widely
implemented Web API practices and those that are still emerging, but which may offer greater
capabilities in the future.
Some organisations (e.g. NOAA and UK Met Office) look to be using Web APIs as a way of bringing
existing legacy systems together and providing a more cohesive API experience for developers.
However these approaches are in their infancy. They do not appear to be using well-defined,
reusable APIs, which we would suggest to be adopted as-is. Yet there are lessons to be learned:
some have identified basic abstractions that may be reusable, such as separation of value-types
(text, image, scalar); observations and forecasts; points, grids, and trajectories, among others.
These may then be used as generic access mechanisms across sub-domains, where such concepts
are generally consistent.
In this report we identify and provide brief overviews of the state of the art in Web API description
languages (5.1), common response types (5.2) and management of Web APIs (5.3). This is an area
of huge growth, following the proliferation of APIs on the Web. Consequently many of these
practices are still immature, so we suggest being cautious when selecting API description
languages and common response types.
The suite of available management tools reflects the need to monitor and manage access to APIs.
These tools appear more mature than service description languages, as there are distinct
commercial opportunities in providing such services. The open source offerings have a range of
functionality out of the box, and appear to be easily adapted for use in particular scenarios.

We provide general recommendations in three areas of provisioning Web APIs:
 General: documentation styles, API presentation, building developer community;
 Technical: use of automated API documentation, API description languages, API managers,
use of the HTTP standard, error and link encoding, versioning, being consistent;
 Modelling the domain: modelling resources, use of domain-driven modelling, involving
domain experts, use of abstractions, and relevant external activities.
These recommendations coupled with the details of existing domain APIs and tools should provide
a basis for more detailed design of environmental Web APIs.

1 Introduction
Environmental monitoring agencies deliver huge value through data delivery. From primary
industries and infrastructure to the general public, there are many consumers of environmental
data. Industries such as aviation, agriculture, emergency management, mining, defence, planning,
and health all make use of environmental data. People access environmental data through a
variety of channels but delivery is increasingly driven by software. Whether powering websites,
social media, email, or publications, data must be timely, reliable and fit for use.
Increasingly organisations are providing web-based access to data to allow development of
innovative applications and third-party products and services. This report reviews current
practices of Web-based Application Programming Interfaces (APIs) for delivery of environmental
data. Also referred to as Web services, APIs are increasingly being used for data delivery through
multiple media channels. We focus on practices that emerged through the ‘Web 2.0’ progression
of Web technologies. These technologies addressed rapid delivery of data to Web-based
applications, such as Facebook, Twitter, and so on. These approaches have strongly influenced
data delivery on the modern Web.
All environmental organisations have an opportunity to improve their ability to assist industries
and drive creation of new businesses by providing clear and descriptive access to its data through
Web APIs. In this report we cover four distinct aspects of the problem, as shown in Figure 1.
Nature of Related
environmental domain Web
data APIs
The need for Web API best

APIs practices
Environmental
data Web APIs
Figure 1 - Method for this report
The structure follows from this: Section 2 defines what Web APIs are and why they are needed;
Section 3 provides a brief overview of the challenges associated with environmental data; Section

4 reviews how other environmental agencies are using Web APIs for data delivery; Section 5
provides a summary of associated Web API practices, which are independent of the specific
domain of application; and Section 5 concludes the report.
1.1 Scope
While this report originates within a hydrology-specific activity (WIRADA), it takes a broader look
at publishing environmental data on the Web. The report does not target any specific areas, but
attempts to look across a range of areas within environmental monitoring and data acquisition.
This report does not review existing standards of the Open Geospatial Consortium (OGC), as these
have been covered in other activities (Swain et al., 2015)(Percivall, 2010) and within the OGC test
beds. We do touch on a number of relevant activities related to the OGC (e.g. the GeoServices
API). Additionally the report does not investigate vocabulary (e.g. code lists) management and
delivery services. We also do no provide an in-depth review of Linked Data services (Heath and
Bizer, 2011), but acknowledge there is influence occurring between Linked Data and Web APIs
(see following section).

2 The need for Web APIs
The term ‘Web API’ has no widely agreed formal definition. It generally refers to a server-side
Application Programming Interface (API). Such APIs provide software interfaces for message
systems based on a request-response pattern.
For the purposes of this report, the following statements apply to the term Web API:
• Mostly associated with the REpresentational State Transfer (REST) architectural style;
• Mostly use JSON encodings;
• The focus tends to be on web and mobile developers as consumers.
• There tends to be less focus on service-to-service interaction, and more on serving web
applications directly.
While Web APIs can be described using the more traditional term ‘Web Services’, these are often
associated with the suite of World Wide Web Consortium (W3C) Web Services standards. In this
sense, ‘Web Services’ typically use some or all of the following standards:
• Simple Object Access Protocol (SOAP);
• Web Service Definition Language (WSDL);
• Universal Description Discovery and Integration (UDDI);
• XML encodings.
The two terms also usually reflect a different user base. Table 1 provides a summary of service
types and related technologies, their typical users and origin. While this report focuses on the
‘Web 2.0’ style of APIs, there appears to be increasing convergence and cross-pollination of ideas
between Linked Data and Web APIs.
Table 1 - Service type categorisation
Primary usage Technologies utilised Core attributes

domain/origin
Enterprise computing XML (XSD, XSLT), W3C  Supports integration of different

Web Services, SOAP, business platforms
enterprise service bus,
 Testable information models with
relational databases.
conformance checks
 Quality of service mechanisms
directly supported
 Detailed metadata
 Strict validation of content

Web 2.0 Javascript (client +  Supports development of dynamic
server-side), HTTP, web applications
JSON, document and
 Loose validation of content
in-memory databases
 Simplified information models
Research Semantic Web, Linked  Interlinked data across the web

Data, RDF
 Web scale data integration
 Distributed, potentially unstable,
infrastructure
 Open, extensible, content models
 Validation is an ongoing challenge1
2.1 Why provide APIs?

Web APIs are becoming crucial to any company operating on the Internet. The growth in number
of APIs provides some indication of how important they are becoming. Figure 2 shows the growth
in APIs registered on http://www.programmableweb.com/.
1
http://www.w3.org/2014/data-shapes/charter

Figure 2 - Growth in APIs (image supplied with permission from ProgrammableWeb.com)
Companies are using APIs as platforms for building their own services, as well as building a
community around their services. It is becoming an engagement model with their customers that
create positive feedback into their businesses.
When a company provides an API, it is providing a service to the developer community. This
involves making data available in some form. Designing Web APIs should be about minimising
friction for developers. This involves carefully balancing complexity with functionality. If a balance
is achieved then developers will use the API, build applications, tools, and write documentation.
An ecosystem begins to form. This greatly benefits the service provider as such an ecosystem
exploits the inherent value in data. It is also begins a feedback loop to services provided by the
service provider.
There are many reasons environmental organisations should provide APIs, including:
 Build mobile apps more quickly. Having a managed API that provides easy access to data
lowers the entry for developers of mobile applications. The application may be developed
externally or internally; the API is part of the platform.
 Customers increasingly want direct access to environmental data. Environmental data is
used in a myriad of ways. And the demand is only increasing. From agricultural services
providing decision support to farmers, to aquaculture companies managing fish pens,
environmental data is crucial to making informed decisions.
 Avoid screen-scraping. Organisations typically provide many web sites with HTML data
tables, and graphics showing data as graphs and maps. Developers sometimes try to get
the underlying data by ‘scraping’ the web-page, sometimes even involving inferring data
from graphics. This is a brittle way to share data. Changes to websites break software;

there is no visibility on how people are accessing the data, or clear methods on controlling
their access.
 Works for external and internal users. Many companies discover that once a solid API is
put in place it becomes equally important for users within the company. This greatly
increases the user base and momentum of the API.
2.2 Who are the users of Environmental Web APIs?

The direct users of Web APIs are software developers. But software is not the end game – it’s the
service provided by the software that creates value. So while developers interact with the API,
their requirements differ according to the driver behind writing the software or application. This is
illustrated in Figure 3.
Figure 3 - API value chain
Understanding the user base (end users) gives us an idea of the requirements for specific
applications. Some of these requirements we may know already (applications exist), some may
have been requested (indicates a need), others we may never predict (novel uses of data).
Describing the potential user groups is a useful starting point for understanding requirements.
A report for the Australian Water Resource Information System (AWRIS) (National Water
Commission, 2006) includes an initial analysis of user requirements for hydrological data. The
scope of this report is wider than hydrological data, however the AWRIS report serves as a useful
starting point.

2.3 Supporting new data-driven businesses
An important part of providing Web APIs across a range of products is the creation of ‘ecosystems’
for business to use environmental data. While building an application on a single product is useful,
there is huge potential in the combination of APIs with other third parties. This can assist in
creation of new business models.
The AWRIS user requirements report describes example scenarios that are used to define
requirements. We reuse the ‘Blue Sky Industries’ fictional scenario - reproduced in the breakout
box below - to explore how APIs can support different businesses through different project
phases.
Blue Sky Industries

Blue Sky Industries are hoping to build a new chemical production plant, and are targeting a
number of locations across the globe, including Australia. They are looking to identify regions
that would be best suited for this type of plant. Prior to detailed investigation, Blue Sky
Industries have developed a set of criteria they hope will provide a short list of suitable
regions. One of the key criteria is access to significant amounts of suitable water for
production. Blue Sky Industries need to know where water is being used in a similar way,
where water is available for use, and if the quality of the available water is suitable for their
requirements.
The bold terms in the scenario description identify data requirements. We further break the
scenario into two phases: an analysis phase focused on viability of the plant, and an operational
phase once a plant is in place. It would be sensible for the company consider other non-water data
sources. Figure 4 describes a fictional way in which different Web APIs could support these two
phases. Third parties are introduced to identify the role of data brokers/service providers. These
providers may offer analytical services to Blue Sky to support the project phases. Blue Sky or the
originating organisation may even provide this role. A scenario like this also provides interesting
options for cost structures for data, analytics and support.

Figure 4 - Scenario data requirements across two phases
The following environmental data products have potential for use in such a scenario:
Analysis phase
 Seasonal forecasts – what are the mid-term expectations around water availability?
 Climate statistics – how can previous conditions inform analysis of the plant viability?
 Water quality – does the water quality match plant operation requirements?
 Catchments/spatial regions – analysis of suitable regions would require an understanding
of catchment areas and general spatial features (nearby assets, hydrological features etc.).
Operational phase
 Short-term weather – monitoring for severe weather events;
 Flow predictions – manage quantities for intakes;
 Seasonal forecasts – medium-term production forecasts based on access to water;
 Water quality – ensure adequate water quality is being maintained.
These are fictional usages of data, but most already exist in industry in some form. The scenario is
used to indicate the new ways in which organisations are becoming more data-driven. Increasing
the ways and ease in which organisations can access data is an important catalyst in the creation
of these businesses.

2.4 What are the challenges for environmental data Web APIs?
Conway’s Law2 afflicts many organisations:
“organizations which design systems ... are constrained to produce designs which are copies of
the communication structures of these organizations”
This is typical for organisations with long histories - there are systems that are isolated and not
connected to other relevant parts. The result is data access over the Internet is segregated: there
are multiple ways to get to similar data, there is overlap between reporting services, and it’s
difficult for developers to find where best to access data.
Additionally, environmental agencies deal in data that are inherently complex. Providing a single,
simple Web API to access data is not trivial. We cannot simply follow the ‘develop a REST service
for your products’ tutorial. Environmental data are diverse and dynamic. From a simple analysis of
a typical environmental agency’s data products and services, we identified 63 distinct Internet
offerings. These range from observations to forecasts, weather to hydrology, and point data to
gridded data.
This leads to the primary question explored in this report: how do organisations provide
consistent Web APIs across a range of environmental data products?
2
https://en.wikipedia.org/wiki/Conway%27s_law

3 The nature of environmental data
Environmental data suffers from all the Big Data challenges, known as the ‘Vs’ of Big Data. This
provides a useful frame to describe the challenges environmental agencies face in handling its
data. Figure 5 provides some examples of how environmental data relate to each of the ‘V’
challenges, which are discussed below.
Organisational complexity
Variety Volume Velocity Veracity
Observations, Cheaper more

Model
forecasts, abundant Data Uncertainty
computation
warnings sensors
Points, grids, Push-based

Data storage Communication
coverages communications
Images, video, Earth Warnings and

Tailored data
scalar, binary observations alerts
Meteorology, Synchronisation
Climatalogy, with other Data validity
Hydrology organisations
Provenance, data Handling 3rd

transparancy party data
Organisational
differences
Figure 5 - Complexities in environmental data
Variety – There is huge variety in the data managed by environmental agencies. Issues requiring
consideration include:
 Temporal nature: observations and forecasts;
 Structure: point and gridded;
 Multi-modal: images, video, spectra, point clouds etc.
 Domain specific: meteorological, hydrological, climatological, oceanography, etc.

 Organisational: use of specific codes, vocabularies, methods, semantics and systems.
 Provenance: tracking data lineage and post-processing to provide transparency to data
users.
Volume – Environmental models can generate huge amounts of data. Models are also coupled
with other models and data sets that are equally large. For example, satellite data that are inputs
to global circulation models that are inputs to regional meteorological models.
Velocity – Data moves quicker than ever. Sensor communications are switching from traditional,
pull-based mechanisms, to near real-time push-based communications. The temporal and spatial
resolution of models is increasing as compute power is able to handle higher loads.
Interdependencies of data moving at different rates create a complex data management
environment. The speed at which people, and machines, demand data has also increased. User’s
expectations around the timeliness of data have increased dramatically over the last 10 years.
Veracity – Introduced later in the definition of Big Data challenges, Veracity – or uncertainty of
data – is a big issue for environmental data. From issues of positioning, and calibration of sensors,
to the inherent uncertainty in any forecast, environmental agencies are in the business of handling
uncertainty. Communication of quality and uncertainty to users of data is a huge challenge, and
one that has a large social component.
Overlaid across these are the organisational and social complexities that exist in any organisation.
Differing data management approaches, systems, language and culture all contribute to
hampering approaches to cohesive data management and publication.

4 Related domain APIs
There is an enormous number of APIs on the Web. ProgrammableWeb3 lists over 14,000 available
APIs in their catalog of APIs. These include APIs for sports, social, weather, finance, design and so
on. Many of these APIs are location or context specific – for example providing an API to Norway’s
library catalog, or an event search API for Europe.
There are environmental APIs within this catalog. Many are registrations of federal/national data
services, such as NOAA, USGS for the US. There has been enormous growth in weather prediction
services, and 182 APIs listed are associated with weather data. In turn, this has spawned services
that aggregate and compare forecast quality across services4.
Reviewing all these APIs was not feasible for this analysis. In this section we cover some of the
main environmental data APIs we have been able to identify. The focus is on environmental
agencies operating meteorological, hydrological, oceanographic, and related monitoring and
forecasting programs. We seek to categorise the key characteristics of these APIs with the aim of
describing the current practices, and direction that Web APIs are heading. In Appendix A we
provide a summary of the APIs and their key characteristics.
4.1 National Oceanic and Atmospheric Administration (NOAA)

The US National Oceanic and Atmospheric Administration (NOAA) contains six organisations that
cover the following areas:
 Environmental Satellite, Data, and Information Service
 Marine Fisheries Service
 Ocean Service
 Weather Service
 Marine & Aviation Operations
 Oceanic and Atmospheric Research
The primary role for data access falls within the Environmental Satellite, Data, and Information
Service. It previously hosted three data centres that are now being brought together into the
National Centers for Environmental Information (NCEI). This includes the (National) Climatic Data
Center, the Geophysical Data Center, and the Oceanographic Data Center.
There are attempts at bringing data and services together. The primary example of this is
https://data.noaa.gov/, which is a CKAN-driven (see section 4.11) data discovery and access portal.
3 http://ww.programmableweb.com
4
http://www.weatherapi.net/compare-forecasts/

This is still under development. However, it acts as a catalog and data upload service rather than a
specific data access point – the URLs for data access are submitted as part of registering a dataset.
The NCEI covers many categories of data as shown in Figure 6.
Figure 6 - NOAA NCEI data categories
Most service links lead to a specific data centre’s web site and associated web service, which,
understandably, shows integration is a long-term goal. There is a large number of APIs available; a
selection from some of the categories are shown in Table 2. The Climate API v2 provides a RESTful
API and is summarised in the following section.
Table 2 - Selection of NOAA APIs
ONLINE APIS DESCRIPTION FORMAT(S)
Climate Data API v2 Search and access datasets and station and/or location JSON
data.
Climate API Legacy Station time-series data. XML, CSV
Climate REST Services Station network locations and climatological GIS JSON, XML
products.
Catalog Services A catalog service that conforms to the HTTP protocol XML
(CSW) binding of the OpenGIS Catalog Service ISO Metadata
Application Profile specification.
Gridded data service Satellite/Radar/Model Data. NetCDF, XML,

NCEI TDS CSV

Numerical model Numerical Model Data. NetCDF, XML,
gridded service: CSV
NOMADS TDS
Severe Weather Data NEXRAD Storm Cell Attributes, Storm Reports. XML, CSV, KMZ,
Inventory Shapefile
Historic Observing Integrated station history database providing in situ JSON

Metadata Repository metadata.
Web Services
4.1.1 Climate Data API v2.0
The basic resources of the NOAA Climate API are summarised in Figure 7.
•Top level grouping

/datasets •‘Annual summaries’, ‘hourly precip’
•A logical grouping of data types

/datacategories •‘Sky cover & clouds’, ‘Evaporation’
•The instance phenomenon

/datatypes •‘Long-term averages of annual growing degree days with base 45F’
•Logical grouping of location types

/locationcategories •‘Hydrologic Region’, ‘Climate Division’
•Individual (point) locations

/locations •Individual US states, cities
•Monitoring stations
/stations •Individual automatic weather station
•Give me the data already

/data •Fetch data from the Daily Summaries for zip code 28801, May 1st of 2010
Figure 7 - Climate data online RESTful API overview
The API provides a landing page (Code listing 1) that describes the available resource end points,
available parameters and whether they are required. This is good practice and generally
recommended within RESTful APIs.

Code listing 1 - Landing page for Climate API v2
[
{
"path": "v2/lcd/availablereporttypes",
"method": "GET",
"parameters": [
{
"name": "stationid",
"required": false
},
{
"name": "locationid",
"required": false
},
{
"name": "startdate",
"required": true
},
{
"name": "enddate",
"required": true
}
]
},
The resources in this API do not make use of embedded links within the JSON response
documents, which would allow automated traversal of the API. It also requires the client
understand the relationship between relevant resources. For example to find precipitation data,
one must use the ‘datatype’ resource to retrieve a list of all the codes; see example HTTP request
and JSON response in Listing 2.
Listing 2 - example datatype listing
GET http://www.ncdc.noaa.gov/cdo-web/api/v2/datatypes?limit=1000 (auth header required)
{
"metadata": {
"resultset": {
"offset": 1,
"count": 1461,
"limit": 1000
}
},
"results": [
{
"mindate": "1994-03-19",
"maxdate": "1996-05-28",
"name": "Average cloudiness midnight to midnight from 30-second ceilometer data (percent)",
"datacoverage": 1,
"id": "ACMC"
},
{
"mindate": "1965-01-01",
"maxdate": "2005-12-31",
"name": "Average cloudiness midnight to midnight from manual observations (percent)",
"datacoverage": 1,
"id": "ACMH"
},
{
"mindate": "1994-02-01",
"maxdate": "1996-05-28",
"name": "Average cloudiness sunrise to sunset from 30-second ceilometer data (percent)",
"datacoverage": 1,
"id": "ACSC"
},

There are many (1461) data types, some with very specific definitions such as ‘Long-term averages
of number of days during December-February with precipitation >= 0.50 inches’ and ‘50th
percentiles of daily nonzero snow depth for 29-day windows centred on each day of the year’.
There are ways of filtering, e.g. return a list of stations that support a specific datacategoryid (see
Listing 3), but this requires the code identifier for the category ID.
Handling code lists and definitions is a common problem in Web APIs. There are some approaches
typically used that can help: the service may provide embedded links to allow direct navigation to
the definition, or provide a definition object with basic properties, plus links for further details.
These approaches can assist by reducing the amount of manual lookups required from clients.
There is increasing interest in the use of vocabulary services to assist in the publication and
management of vocabularies (Cox et al., 2014; Schandl and Blumauer, 2010; Yu et al., 2011).
Review of such services is out of scope for this report. However, it is recommended that such
services be investigated when deploying Web APIs that use code lists and/or structured
vocabularies.
Listing 3 – query stations that support specific datatype
GET http://www.ncdc.noaa.gov/cdo-web/api/v2/stations?datatypeid=ANN-PRCP-AVGNDS-GE001HI (auth

header required)
{
"metadata": {
"resultset": {
"offset": 1,
"count": 7480,
"limit": 25
}
},
"results": [
{
"elevation": 408.4,
"mindate": "1980-01-01",
"maxdate": "2012-08-31",
"latitude": -14.31667,
"name": "AASUFOU, US",
"datacoverage": 1,
"id": "GHCND:AQC00914000",
"elevationUnit": "METERS",
"longitude": -170.76667
},
{
"elevation": 3.7,
"mindate": "1945-08-01",
"maxdate": "2015-12-30",
"latitude": -14.33056,
"name": "PAGO PAGO WEATHER SERVICE OFFICE AIRPORT, US",
"datacoverage": 1,
"id": "GHCND:AQW00061705",
"elevationUnit": "METERS",
"longitude": -170.71361
},
In addition to the datatype ID, a dataset ID is required, in the example case this is requested as
shown in Listing 4.
Listing 4 - requesting a dataset for a specific datatype
GET http://www.ncdc.noaa.gov/cdo-web/api/v2/datasets?datatypeid=ANN-PRCP-AVGNDS-GE001HI (auth

header required)

{
"metadata": {
"resultset": {
"offset": 1,
"count": 1,
"limit": 25
}
},
"results": [
{
"uid": "gov.noaa.ncdc:C00821",
"mindate": "2010-01-01",
"maxdate": "2010-01-01",
"name": "Normals Annual/Seasonal",
"datacoverage": 1,
"id": "NORMAL_ANN"
}
]
}
This then allows a query to be made against the ‘data’ resource itself, as shown in Listing 5.
Listing 5 - requesting data for multiple stations of a particular datatype, within a time window
GET http://www.ncdc.noaa.gov/cdo-web/api/v2/data?datatypeid=ANN-PRCP-AVGNDS-
GE001HI&startdate=2010-01-01&enddate=2010-02-06&datasetid=NORMAL_ANN (auth header required)
{
"metadata": {
"resultset": {
"offset": 1,
"count": 7484,
"limit": 25
}
},
"results": [
{
"date": "2010-01-01T00:00:00",
"datatype": "ANN-PRCP-AVGNDS-GE001HI",
"station": "GHCND:AQC00914000",
"attributes": "P",
"value": 2275
},
{
"date": "2010-01-01T00:00:00",
"datatype": "ANN-PRCP-AVGNDS-GE001HI",
"station": "GHCND:AQW00061705",
"attributes": "C",
"value": 2495
},
continues for 7848 points
The use of disjoint resources allows query flexibility, but also requires the client to understand the
resource model and which queries are logical. This is done through a combination of reading
documentation and making example queries.
Key characteristics:
 Incomplete metadata.
 Uses code lists without much description of code specifics.
 No use of links/hypermedia.

 Implicit CRS for lat-lon pairs.
4.2 UK Met Office – DataPoint

The UK Met Office runs a beta RESTful API called DataPoint5, which provides access to
observations and forecasts, using JSON, XML and imagery and text formats. The API generalises
along three concepts: the product, data type and content type. This is summarised in Figure 8.
Content type Data type Product
• value • observation • mountainarea

• image • forecast • surfacepressure
• layer • ukextremes
• text • nationalpark
• all
Figure 8 - DataPoint API overview
The service does not provide links and thus lacks the hypermedia component of RESTful APIs. The
JSON representation appears to be a direct translation from XML, and thus mimics XML’s
structure. This in itself is not a problem, however in some cases it maps attributes using an
ampersand character (e.g. “@name”), which is OK for plain JSON but may conflict in Linked Data
JSON (JSON-LD) where ampersands are used for semantic mark-up.
The capabilities requests provide a summary of available datasets within a particular product; see
Listing 6 for an example listing of forecast layers. An example site listing response is shown in
Listing 7. There is no paging applied to this response, which results in all sites being returned
(1.4MB JSON response). Listing 8 shows an example of a regional text forecast encoded in JSON.
Listing 6 - example capabilities for forecast layers (modified for formatting)
GET
http://datapoint.metoffice.gov.uk/public/data/layer/wxfcs/all/json/capabilities?res=hourly&key=<K
EY>
5
http://www.metoffice.gov.uk/datapoint

{
"Layers": {
"@type": "Forecast",
"BaseUrl": {
"@forServiceTimeFormat": "Timesteps",
"$":
"http://datapoint.metoffice.gov.uk/public/data/layer/wxfcs/{LayerName}/{ImageFormat}?RUN={DefaultTime}Z
&FORECAST={Timestep}&key={key}"
},
"Layer": [
{
"@displayName": "Rainfall",
"Service": {
"@name": "UKPPBEST",
"LayerName": "Precipitation_Rate",
"ImageFormat": "png",
"Timesteps": {
"@defaultTime": "2016-01-04T21:00:00",
"Timestep": [
0,3,6,9,12,15,18,21,24,27,30,33,36
]
}
}
},
{
"@displayName": "Cloud",
"Service": {
"LayerName": "Total_Cloud_Cover",
"Timesteps": {
"@defaultTime": "2016-01-04T21:00:00",
"Timestep": [
0,3,6,9,12,15,18,21,24,27,30,33,36
]
}
}
},
{
"@displayName": "CloudAndRain",
"Service": {
"LayerName": "Total_Cloud_Cover_Precip_Rate_Overlaid",
"Timesteps": {
"@defaultTime": "2016-01-04T21:00:00",
"Timestep": [
0,3,6,9,12,15,18,21,24,27,30,33,36
]
}
}
},
{
"@displayName": "Temperature",
"Service": {
"LayerName": "Temperature",
"Timesteps": {
"@defaultTime": "2016-01-04T21:00:00",
"Timestep": [
0,3,6,9,12,15,18,21,24,27,30,33,36
]
}
}
}
<snipped>
}
]
}

Listing 7 – Example location listing in JSON
GET http://datapoint.metoffice.gov.uk/public/data/val/wxfcs/all/json/sitelist?key=<KEY>
{
"Locations": {
"Location": [
{
"elevation": "933.0",
"id": "3072",
"latitude": "56.879",
"longitude": "-3.42",
"name": "Cairnwell",
"nationalPark": "Cairngorms National Park",
"region": "ta",
"unitaryAuthArea": "Perth and Kinross"
},
{
"elevation": "134.0",
"id": "3088",
"latitude": "56.852",
"name": "Inverbervie",
"region": "gr",
"unitaryAuthArea": "Aberdeenshire"
},
{
"elevation": "4.0",
"id": "3094",
"latitude": "57.698",
"name": "Rosehearty Samos",
"region": "gr",
"unitaryAuthArea": "Aberdeenshire"
}
…continues
Listing 8 - example text regional forecast, encoded in JSON
GET http://datapoint.metoffice.gov.uk/public/data/txt/wxfcs/regionalforecast/json/507?key=<KEY>

{
"RegionalFcst": {
"createdOn": "2016-01-04T13:11:06",
"issuedAt": "2016-01-04T16:00:00",
"regionId": "nw",
"FcstPeriods": {
"Period": [
{
"id": "day1to2",
"Paragraph": [
{
"title": "Headline:",
"$": "Rather cloudy with occasional showers,"
},
{
"title": "This Evening and Tonight:",
"$": "This evening and tonight will be cloudy with showery outbreaks of rain which could
be heavy at times. The showers could merge to give some longer spells of rain, particularly over the
Pennines. Minimum Temperature 5C."
},
{
"title": "Tuesday:",
"$": "Tuesday will be another rather cloudy day with further showery outbreaks of rain
which could be heavy at times. Maximum Temperature 8C."
}
]
},
{
"id": "day3to5",
"Paragraph": {
"title": "Outlook for Wednesday to Friday:",
"$": "Mainly cloudy but drier on Wednesday. Heavy rain on and strong winds Wednesday night,
clearing to sunshine and heavy showers early on Thursday, with further heavy showers on Friday."
}
},
{
"id": "day6to15",
"Paragraph": {
"title": "UK Outlook for Saturday 9 Jan 2016 to Monday 18 Jan 2016:",
"$": "The weekend will be unsettled and often windy with showery conditions interspersed
with some more organised spells of rain, and snow on northern hills. The heaviest rain and strongest
winds occurring in the southwest and the northeast, where large rainfall totals and gales are expected.
Next week will continue unsettled, with further showers or longer spells of rain, interspersed by some
brief dry and brighter periods. It will become windy at times, particularly at the start of next week
when severe gales are possible. Temperatures through the period are likely to be near or a little above
average. However, there is the risk of a brief much colder spell at the end of next week, with frost
and snow showers."
}
},
{
"id": "day16to30",
"Paragraph": {
"title": "UK Outlook for Tuesday 19 Jan 2016 to Tuesday 2 Feb 2016:",
"$": "A generally unsettled spell of weather is expected through this period with winds
mainly coming from the west. Showers or longer spells of rain are expected, interspersed by some dry
and brighter interludes. The rain being heaviest and most persistent in the west where rainfall totals
are likely to be above average. The east should be more sheltered, getting the best of the drier and
brighter interludes, so rainfall totals here may by slightly below average. There are indications of
some longer dry spells later in January, especially in the south. Temperatures through this period
should be near or a little above average for most areas."
}
}
]
}
}
}
Listing 9 - example text observations of UK extremes (showing one region only)
GET http://datapoint.metoffice.gov.uk/public/data/txt/wxobs/ukextremes/json/latest?key=<KEY>

{
"UkExtremes": {
"extremeDate": "2016-01-04",
"issuedAt": "2016-01-04T23:03:09Z",
"Regions": {
"Region": [
{
"id": "sw",
"name": "South West England",
"Extremes": {
"Extreme": [
{
"locId": "03853",
"locationName": "Yeovilton",
"type": "HMAXT",
"uom": "degC",
"$": "11.0"
},
{
"locId": "03710",
"locationName": "Liscombe",
"type": "LMAXT",
"uom": "degC",
"$": "6.6"
},
{
"locId": "03647",
"locationName": "Little Rissington",
"type": "LMINT",
"uom": "degC",
"$": "3.1"
},
{
"locId": "03823",
"locationName": "Cardinham",
"type": "HRAIN",
"uom": "mm",
"$": "23.2"
},
{
"locId": "03853",
"locationName": "Yeovilton",
"type": "HSUN",
"uom": "hours",
"$": "2.2"
}
]
}
}
 Observations and forecasts supported
 Capabilities for discovery of available time slices
 Text-based documentation
 No linking
 JSON or XML responses available
 Supports imagery and descriptive text outputs
 Implicit CRS for lat/lon values

4.3 Forecast.io
Forecast.io provides an API6 to access its aggregated forecast data. This is used to drive its mobile
application DarkSky7. The API it provides is a simple RESTful service that provides JSON encoded
data of:
 Current conditions
 Minute-by-minute forecasts out to 1 hour (where available)
 Hour-by-hour forecasts out to 48 hours
 Day-by-day forecasts out to 7 days.
The API is divided into two main calls: current forecast for next 7 days at a specific location (lat,
lon), and observed or forecast conditions for a specific location (lat, lon) and time (60 years into
the past to 10 years into the future).
The response structure is simple: arrays of data blocks containing both hourly and daily data
points. Each data point contains a UNIX time with associated forecast values. See listing below for
example data point structure.
{
"time": 1448283600,
"summary": "Mostly cloudy throughout the day.",
"icon": "partly-cloudy-day",
"sunriseTime": 1448303752,
"sunsetTime": 1448356686,
"moonPhase": 0.43,
"precipIntensity": 0.0152,
"precipIntensityMax": 0.0381,
"precipIntensityMaxTime": 1448344800,
"precipProbability": 0.03,
"precipType": "rain",
"temperatureMin": 11.63,
"temperatureMinTime": 1448298000,
"temperatureMax": 17.55,
"temperatureMaxTime": 1448334000,
"apparentTemperatureMin": 11.63,
"apparentTemperatureMinTime": 1448298000,
"apparentTemperatureMax": 17.55,
"apparentTemperatureMaxTime": 1448334000,
"dewPoint": 8.92,
"humidity": 0.71,
"windSpeed": 6.49,
"windBearing": 283,
"cloudCover": 0.61,
"pressure": 1008.45,
"ozone": 305.7
}
6 https://developer.forecast.io/docs/v2
7
http://darkskyapp.com/

 Use of UNIX time in responses. ISO8601 are available for time-based requests.
 All values are JSON numbers with no metadata. All metadata resides in the documentation.
E.g. ‘pressure: A numerical value representing the sea-level air pressure in millibars.’
 Units are grouped using a ‘unit’ query parameter. This makes it possible to request SI units
using a query parameter and all units are translated. There is an ‘auto’ request, which will
lookup the units associated with geolocation specified.
 Latitude and longitude are assumed to be in a (non-specified) hardcoded CRS.
4.4 USGS Water Data Services

The USGS offer a range of Web APIs8 for water data including a site service, daily values service,
groundwater service, water quality service, instantaneous values service, and a statistics service.
These services tend to follow a key-value pair style of API. The documentation is not generated
from the code interfaces, but very precise in its definition of available parameters and queries.
Test tools are available for all services, which provide an HTML form-based tool for creating
queries to the services.
The services support multiple output formats (XML, JSON, CSV), along with support for some data
transfer standards, e.g. WaterML2.0. The services are consistent in their presentation and use of
common practices.
The JSON response format appears to be a mapping of the WaterML1.1 XML structure; see Listing
10 for an example response from the instantaneous value service. WaterML1.1 is non-OGC version
of WaterML that didn’t make use of existing standards, such as Observations & Measurements.
Listing 10 - example instantaneous value request in JSON
GET
http://waterservices.usgs.gov/nwis/iv/?format=json,1.1&sites=01646500&parameterCd=00060,00065
8
http://waterservices.usgs.gov/rest/

{
"name": "ns1:timeSeriesResponseType",
"declaredType": "org.cuahsi.waterml.TimeSeriesResponseType",
"scope": "javax.xml.bind.JAXBElement$GlobalScope",
"value": {
"timeSeries": [{
"sourceInfo": {
"siteName": "POTOMAC RIVER NEAR WASH, DC LITTLE FALLS PUMP STA",
"siteCode": [{
"value": "01646500",
"network": "NWIS",
"siteID": null,
"agencyCode": "USGS",
"agencyName": null,
"default": null
}],
"timeZoneInfo": {
"defaultTimeZone": {
"zoneOffset": "-05:00",
"zoneAbbreviation": "EST"
},
"daylightSavingsTimeZone": {
"zoneOffset": "-04:00",
"zoneAbbreviation": "EDT"
},
"siteUsesDaylightSavingsTime": false
},
"geoLocation": {
"geogLocation": {
"srs": "EPSG:4326",
"latitude": 38.94977778,
"longitude": -77.12763889
},
"localSiteXY": []
},
"elevationM": null,
"verticalDatum": null,
"note": [],
"extension": null,
"altname": null,
"siteType": [],
},
"variable": {
"variableCode": [{
"value": "00060",
"network": "NWIS",
"vocabulary": "NWIS:UnitValues",
"variableID": 45807197,
"default": true
}],
"variableName": "Streamflow, ft³/s",
"variableDescription": "Discharge, cubic feet per second",
"valueType": "Derived Value",
"dataType": null,
"generalCategory": null,
"sampleMedium": null,
"unit": {
"unitName": null,
"unitDescription": null,
"unitType": null,
"unitAbbreviation": "ft3/s",
"unitCode": null,
"unitID": null
},
continues…
The USGS has also recently introduced a statistics web service and is similar in design (key-value
pairs). It provides daily, monthly and annual statistics for continuous monitoring stations. It
currently only supports CSV responses, with JSON and XML planned for future work.

 Detailed documentation.
 Multiple services, with general consistent behaviour and practices (across the water data
services).
 Support for standard encodings.
 Key-value pair style API.
 Separation of services by temporal dimension (e.g. instantaneous and daily) and sub-
domain (surface water, groundwater, and water quality).
 No hypermedia/links in responses.
 Explicit CRS.
4.5 Atlas of Living Australia

The Atlas of Living (ALA) has a comprehensive API for accessing species data, including
occurrences, taxonomies, spatial data (distributions, points of interest etc.) and plots. It provides
both output and input options to support crowd-sourced data.
Service type Description Example queries
Species Profile Taxonomic name, concept Search for all ‘Acacia’ species
lookups, autocomplete
services.
Occurrence Specimen & observation data Search for the records for the
searching. genus Macropus returning
taxonomic, geospatial,
temporal information for
occurrences.
Geospatial Get/put spatial layers. Download a shape object as

GeoJSON using its primary ID.
Get/put points of interest.
Get/put geometries.
Including intersection services
and gazetteer information.
Mapping Creating maps with WMS Generates a static density heat

services, static heat maps. map in PNG format.
Endemism Services for reports on List endemic species within an

endemism for an area. area.

Collections Collections metadata including List details of Australian
taxonomic scope, attribution. National Insect Collection.
Data Resources Data resource metadata Get metadata associated with

including taxonomic scope, a species list.
attribution.
Data Providers Data provider metadata Get data from the CSIRO
including taxonomic scope, National Fish Collection.
attribution.
Institution Institution metadata including Get details of the Tasmanian

taxonomic scope, attribution. Museum and Art Gallery.
Crowd sourcing Crowd sourced data – Retrieve number of user

contributions, validation, contributions by month.
automated harvesting, stats
for user contributions.
General data Dashboard data feeds, data Get list of all licences used by
licences, region lists etc. ALA.
Taxonomy Taxonomic services. E.g. name Higher taxa for Macropus rufus
lookups, higher order lookups.
.Distributions Services for accessing species Display expert distribution map

distribution polygons for Monacanthus chinensis
Annotations Annotations, assertions for List all assertion codes used.

data.
Scatterplots Services for the generation of Example for Macropus Rufus.

scatterplots for occurrence
data.
Species lists Web services for storage, Get a species list for ‘Red
creation and querying of lists. Kangaroo’.
Listing 11 - example listing for an institution
GET http://collections.ala.org.au/ws/institution/in25

{
"name": "Tasmanian Museum and Art Gallery",
"acronym": "TMAG",
"uid": "in25",
"guid": "urn:lsid:biocol.org:col:34362",
"address": {
"street": "Dunn Place",
"city": "Hobart",
"state": "Tasmania",
"postcode": "7000",
"country": "Australia",
"postBox": null
},
"phone": "(03) 6165 7000",
"email": "herbarium@tmag.tas.gov.au",
"pubDescription": "Primarily, the collections are of the Tasmanian fauna.\r\n\r\nThe Tasmanian
Museum and Art Gallery is making a range of species information and data from its natural history
collections available through the Atlas of Living Australia.",
"techDescription": "Some primary types held and stored separately.",
"focus": null,
"latitude": -42.8817500000,
"longitude": 147.3321520000,
"state": "Tasmania",
"websiteUrl": "http://www.tmag.tas.gov.au/",
"alaPublicUrl": "http://collections.ala.org.au/public/show/in25",
"imageRef": {
"filename": "Astacopsis_gouldi_300px.jpg",
"caption": "Tasmanian Giant Freshwater Lobster",
"copyright": null,
"attribution": "Photo courtesy: Tasmanian Museum and Art Gallery",
"uri": "http://collections.ala.org.au/data/institution/Astacopsis_gouldi_300px.jpg"
},
"logoRef": {
"filename": "TasGov_HorC.png",
"caption": "Tasmanian Govt, TMAG logo",
"copyright": null,
"attribution": null,
"uri": "http://collections.ala.org.au/data/institution/TasGov_HorC.png"
},
"networkMembership": [{
"name": "Council of Heads of Australian Faunal Collections",
"acronym": "CHAFC",
"logo": "http://collections.ala.org.au/data/network/CHAFC_sm.jpg"
}, {
"name": "Council of Heads of Australian Entomological Collections",
"acronym": "CHAEC",
"logo": "http://collections.ala.org.au/data/network/chaec-logo.png"
}],
"hubMembership": [{
"uid": "dh1",
"name": "Online Zoological Collections of Australian Museums",
"uri": "http://collections.ala.org.au/ws/dataHub/dh1"
}],
"attributions": [{
"name": "Biodiversity Collections Index",
"url": "http://www.biodiversitycollectionsindex.org"
}, {
"name": "Council of Heads of Australasian Herbaria",
"url": "http://www.chah.gov.au/resources/index.html"
}, {
"name": "Tasmanian Museum and Art Gallery",
"url": "http://www.tmag.tas.gov.au/"
}],
"dateCreated": "2010-08-31T05:52:27Z",
"lastUpdated": "2014-10-28T02:05:53Z",
"userLastModified": "miles.nicholls@csiro.au",
"institutionType": null,
"collections": [{
"name": "Tasmanian Museum and Art Gallery Invertebrate Collection",
"uri": "http://collections.ala.org.au/ws/collection/co198",
"uid": "co198"
}, {
"name": "Tasmanian Museum and Art Gallery Vertebrate Collection",
"uid": "co111"
}, {
"name": "Tasmanian Herbarium",
"uid": "co60"
}],
"parentInstitutions": [],
"childInstitutions": [],
"linkedRecordProviders": [{
Listing 12 - example spatial object request (state/territory)
GET http://spatial.ala.org.au/ws/object/3742602
{
"name": "Australian Capital Territory",
"id": "Australian Capital Territory",
"description": "Australian Capital Territory, Territory",
"pid": "3742602",
"wmsurl":
"http://spatial.ala.org.au/geoserver/wms?service=WMS&version=1.1.0&request=GetMap&layers=ALA:Objects&fo
rmat=image/png&viewparams=s:3742602",
"area_km": 2428.0110849586404,
"bbox": "POLYGON((148.761582 -35.92208,148.761582 -35.1119459999999,150.772781 -
35.1119459999999,150.772781 -35.92208,148.761582 -35.92208))",
"fid": "cl22",
"fieldname": "Australian States and Territories"
}
The ALA documentation also provides some interesting features, such as the ability to view
historical calls to the public API, including frequency, slow requests, error requests etc. (Figure 9).

Figure 9 - Example request frequency and call history for specific API call
 A lot of use of linked resources.
 Good level of metadata.

 Extensive functionality (maps, parsing data, crowd sourced data, autocomplete services
etc.).
 Dynamic documentation with examples.9
 Supports insertion, deletion, and updating of data.
 Implicit CRS.
4.6 OGC SensorThings API

SensorThings is a proposed OGC standard that addresses data access for the Internet of Things
(IoT). The vision of IoT is having devices in the world connected to the Internet to allow data
retrieval and control. Part 1 is the current proposed standard and deals with access to data; part 2
is about control of devices, and is planned for future work.
SensorThings part 1 (referred to as SensorThings from here on in) provides a RESTful, JSON
encoded API to retrieve data and metadata about ‘things’ that generate streams of data. The
underlying information/resource model for the API uses O&M and has been influenced by the
other SWE standards. The API follows patterns defined by the OData protocol 10. Figure 10 shows
an overview of the core resource model.
9 http://api.ala.org.au/
10
http://www.odata.org/

Figure 10 - SensorThings resource model
SensorThings splits the O&M model across two classes: Datastream and Observation. Datastreams
are the top-level class with a subset of the O&M Observation properties: observedProperty,
resultTime, phenomenonTime (see example in Listing 13). The observations property is analogous
to result in the OM_Observation type, but in this case contains a set of Observation objects (see
example in Listing 14). Each of these is typed according to its observation type from O&M
(Measurement, Geometry etc.).
Listing 13 – A SensorThing Datastream
{
"@iot.id": 1,
"@iot.selfLink": "http://example.org/v1.0/Datastreams(1)",
"Thing@iot.navigationLink": "HistoricalLocations(1)/Thing",
"Sensor@iot.navigationLink": "Datastreams(1)/Sensor",
"ObservedProperty@iot.navigationLink": "Datastreams(1)/ObservedProperty",
"Observations@iot.navigationLink": "Datastreams(1)/Observations",
"description": "This is a datastream measuring the temperature in an oven.",
"unitOfMeasurement": {
"name": "degree Celsius",
"symbol": "°C",
"definition": "http://unitsofmeasure.org/ucum.html#para-30"
},
"observationType": "http://www.opengis.net/def/observationType/OGCOM/2.0/OM_Measurement",
"observedArea": {
"type": "Polygon",
"coordinates": [[[100,0],[101,0],[101,1],[100,1],[100,0]]]
},
"phenomenonTime": "2014-03-01T13:00:00Z/2015-05-11T15:30:00Z",
"resultTime": "2014-03-01T13:00:00Z/2015-05-11T15:30:00Z"
}

Listing 14 - Observation
{
"@iot.id": 1,
"@iot.selfLink": "http://example.org/v1.0/Observations(1)",
"FeatureOfInterest@iot.navigationLink": "Observations(1)/FeatureOfInterest",
"Datastream@iot.navigationLink":"Observations(1)/Datastream",
"phenomenonTime": "2014-12-31T11:59:59.00+08:00",
"resultTime": "2014-12-31T11:59:59.00+08:00",
"result": 70.4
}
The structure of the JSON encoding follows patterns defined by OData. It does not strictly follow
the OData specification (Liang et al., 2015, p. 19), but reuses the common patterns, identifier, and
link structures.
 Uses OData as a pattern for response encoding and link handling.
 Generic IoT problem space – applicable to most sensor applications.
 Light metadata.
 Linked resources.
 Implicit CRS.
4.7 UK National Biodiversity Network

The UK National Biodiversity Network (NBN) provides an API similar to that of the ALA, with a
detailed API providing access to sites, taxon information, observations, organisations etc. It
provides XML and JSON responses, with a service description provided in WADL11 (see section
5.1.4).
4.8 GeoServices API

The GeoServices API is a suite of RESTful APIs are a slightly modified version of the ESRI APIs that
are offered as part of the ESRI commercial GIS software packages. They were proposed as OGC
standards in 2013, but subsequently withdrawn from the submission process. They provide a
range of functionality, summarised in Table 3, which have some overlap with existing OGC web
services.
These services encode JSON responses but make very little use of links, and thus lack the
hypermedia attributes of RESTful services.
Table 3 - summary of GeoServices APIs
Service Description Related OGC
11
https://data.nbn.org.uk/Documentation/Web_Services/Web_Services-REST/resources/restapi/application.wadl

services/standards
Core Defines common requirements classes OWS Common; Web Feature

across all services. This includes: Service, Web Mapping Service.
OGC Abstract Topics.
 Common return structures.
 Extensibility mechanisms.
 Use of HTTP headers and return
codes.
 Idempotency.
 JSON-P support (facilities cross-
domain JSON requests).
 Cross-origin resource sharing
support.
 Encoding coordinate reference
systems.
 Geometry encodings.
 Feature encodings, limited to one
geometry property.
 Coded values.
 Symbol encodings.
Catalog Provides a simple catalog of available Catalog Service for the Web
services. (CS/W)
Map Service Access to maps as images, tiled maps, Web Mapping Service (WMS)
maps with coordinate transformations,
time-indexed maps, request/identify
features using specific geometries, search
maps using query parameters.
Feature Service Retrieve, add, update and delete spatial Web Feature Service (WFS)
features. Add attachments to features.
Search relationships between features.
Geometry Service Provides common geometry operations, Web Processing Service (WPS),
including: project, simplify, length and
area calculations, distance, convex hull,
cut, difference, intersect etc.
Image Service Service to retrieve and query imagery Web Coverage Service (WCS),
data (e.g. satellite imagery). Supports Web Mapping Service (WMS).

mosaics and coordinate transformations.
Geoprocessing Service Execute and monitor geoprocessing tasks Web Processing Service (WPS)
asynchronously or synchronously.
Geocoding Service Lookup addresses relating to locations OGC Open Location Services
and visa versa (reverse geocoding). (OpenLS)
 Wide geospatial functionality.
 No use of links.
 No use of existing OGC standards.
 Uses JSON Schema to describe responses.
 Good separation of API endpoints into clear functions.
 Formal, not adopted, document-based specification.
4.9 Global Change Information System

The US Global Change Research Program (USGCRP) established the Global Change Information
System (GCIS) to integrate any information relating to changes in the global environment. This
includes significant datasets, reports, findings, people, models, images etc. It has been developed
following RESTful principles, as well as those of Linked Data.
While its scope is huge, it has made practical first steps towards useful, accessible and linked
datasets. A summary of the key resources within its RESTful API is shown in Figure 11. Each bubble
below the top level (marked black) is a RESTful resource endpoint that provides access to a JSON
encoding of the resource, with links provided to other relevant resources (e.g. links from
publications, to authors and related datasets).

Figure 11 - Overview of the GCIS resource data model
The RESTful API is documented using Swagger (see section 5.1.1) as shown in Figure 12. The
potential information available through is API is vast. Faceted search is available that supports
spatial and thematic drill-down style discovery12 of resources. A SPARQL endpoint is also
available13 for Linked Data clients. Multiple response formats are available to maximise data
availability. For example, an organisation may be viewed as images of linked resources (e.g.
CSIRO14). This facilitates exploration of the related resources, such as exploring reports published
from people within an organisation.
12 https://gcis-search.jpl.net/search
13 http://data.globalchange.gov/sparql
14
http://data.globalchange.gov/organization/commonwealth-scientific-industrial-research-organisation.svg

Figure 12 - GCIS RESTful API documentation
 Expansive scope.
 Swagger-based documentation.
 Lots of use of linking.
 Support for Linked Data encodings and access.
4.10 UK Environmental Agency – Bathing Water Quality API

The Bathing Water Quality provides access to the UK’s water quality sampling data as well as the
resulting assessments (compliance, swimming status etc.). The API has been developed as a Linked

Data API15, which is a specification of how to provide simple RESTful APIs over linked data. A
summary of the scope and structure of the API is shown in Figure 13.
Figure 13 - UK Bathing Water Quality API
The Agency also provides a Bathing Water Profile16 as a way of grouping information relating to
specific bathing locations, such as the beach, rivers streams and pollution sources.
 Detailed documentation.
 Re-use of standard/common vocabularies.
 ‘Native’ Linked Data – supports multiple formats.
 Clear URI guidelines and adherence to LDA guidelines17.
4.11 CKAN
The Comprehensive Knowledge Archive Network (CKAN) is an open source data management
system that supports registration of data assets by third parties. It is used by a range of open data
15 https://code.google.com/p/linked-data-api/wiki/Specification
16 http://environment.data.gov.uk/bwq/doc/api-bwp-reference-v0.1.html#BathingWaterProfile
17
https://github.com/UKGovLD/linked-data-api/blob/wiki/Specification.md

web sites, including http://datahub.io, http://data.gov.uk, http://data.gov.au/ and
http://publicdata.eu/.
The functionality of the base CKAN web application includes search facilities, metadata retrieval
and entry, geospatial functions, social extensions, visualisation and auditing. The RESTful API
provides access to all of these functions.
 Generic data management and publication technology
 RESTful, JSON-based API for creating, updating and deleting records
4.12 Data Access Protocol (DAP), OpenDAP, Hyrax

OPeNDAP is the “Open-source Project for a Network Data Access Protocol”, which provides open
source implementations of the Data Access Protocol (DAP). The primary implementation is called
Hyrax. DAP is an HTTP protocol that provides access to ‘science’ data without any dependence on
how the format in which the data may be stored. It supports four generic data types: grids,
sequences, arrays, and structures.
The DAP protocol uses HTTP GET requests to convey structural metadata (dataset descriptor),
‘semantic’ (dataset attribute) metadata and the data itself. A summary of these message types is
shown in Table 4.
Table 4 - DAP messages
Request Response Content-type

Dataset Descriptor Structure DDS or Error text/plain
(DDS)
Dataset Attribute Structure DAS or Error text/plain
(DAS)
Data Dataset Descriptor DataDDS or Error application/octet
Structure (DataDDS)
Server version Version information as text text/plain
Help Help text describing all request- text/plain
response pairs
These messages allow clients to interrogate data sources and retrieve data in a binary encoding.
Additionally, a constraint expression URL syntax allows name/variable selection (called
projections) and variable filtering against relational expressions (called selections).
An example projection would be accessing an array using a hyperslab (array index) expression, e.g.
[0:4]. An example selection would be a regular expression search across station names.
In terms of Web APIs, the DAP is a lower-level protocol which provides access to common data
structures. For less sophisticated users or client requirements, it is likely that a layer would be
provided on top of DAP that hides some of its complexity. An example of this is ERDDAP (see
section 4.14).

4.13 THREDDS
Thematic Real-time Environmental Distributed Data Services (THREDDS) is a data catalog/server
that provides web-based access to binary data sources, such as NetCDF, HDF, OPeNDAP and
NEXRAD. It publishes a range of service interfaces including Web Coverages Service (WCS) Web
Mapping Service (WMS), OPeNDAP and NetCDF subsetting service. It uses Unidata’s Common Data
Model (CDM)18 as the key abstraction, which is summarised in Figure 14. This model is very similar
to the abstractions used by the DAP.
Figure 14 - Overview of Unidata's Common Data Model (CDM)
THREDDS is not a Web API in the sense of the definition within this report, however it does
provide some REST-like interfaces and is a component within larger aggregated environmental
data system discussed below.
18
http://www.unidata.ucar.edu/software/thredds/current/netcdf-java/CDM/index.html

4.14 ERDDAP
ERDAPP is a data server that provides access to a range of data sources, using a range of formats
and service interfaces. It can be viewed as a middleware data aggregator that supports integration
of common scientific data formats and services. It does this using two abstractions for data
structures: grids and tabular data. It makes use of DAP (see section 4.12) to do this.
ERDAPP provides a RESTful interface for searching and accessing data from the available services.
It works well with THREDDS services and can be used to provide more simplified, consolidated
access points to multiple THREDDS services.
The basic resources defined in the ERDAPP RESTful API are shown in Table 5.
Table 5 - summary of ERDDAP resources
Resource URL
info http://coastwatch.pfeg.noaa.gov/erddap/info/index.htmlTable?page=1&itemsPerP
age=1000
search http://coastwatch.pfeg.noaa.gov/erddap/search/index.htmlTable?page=1&itemsPe
rPage=1000&searchFor=
categorize http://coastwatch.pfeg.noaa.gov/erddap/categorize/index.htmlTable?page=1&item
sPerPage=1000
griddap http://coastwatch.pfeg.noaa.gov/erddap/griddap/index.htmlTable?page=1&itemsP
erPage=1000
tabledap http://coastwatch.pfeg.noaa.gov/erddap/tabledap/index.htmlTable?page=1&items
PerPage=1000
wms http://coastwatch.pfeg.noaa.gov/erddap/wms/index.htmlTable?page=1&itemsPerP
age=1000
Using its key abstraction, ERDAPP can source data from SOS, WCS, WFS, ArcGIS, text, DAP, NetCDF
among others, and republish using KML, JSON, R, WFS, SOS, WCS and so on. In this sense it is a
powerful data server that can operate as a broker between a users’ desired output and the
original data source.
 Aggregator of common data access protocols and standards
 RESTful API using grid and table simplification pattern
 Supports multiple service interfaces and response formats

5 Web API practices
5.1 Describing service interfaces

Having machine-readable descriptions of APIs is a useful feature to allow automated
documentation, testing, discoverability, generated client bindings, and so on. The Web Service
Definition Language (WSDL) is the W3C standard for service descriptions associated with more
traditional Web Services (as defined in section 2), typically using XML.
There has been recent growth in specifications describing RESTful APIs. These arose from the need
to automatically produce API documentation. However, there are other benefits to having
machine-readable description of an API endpoint, including discoverability and support for
automated testing frameworks.
Some of the more popular frameworks and technologies are identified below. These are largely
bottom-up specifications that arose from implementations. It would be desirable to have an open
standard in this space, since there is already strong similarity between these specifications. This
section does not attempt to pick a winner, but provides short summaries of their background and
basic functions.
5.1.1 Swagger specifications
Swagger is a set of tools and services that help API developers generate documentation for APIs.
The Swagger UI19 renders a documentation page that provides full description of endpoints, with
in-line forms allowing test calls to be made to services. The UI can be generated from an API
specification document, which could be handwritten or generated from a service.
The specification of a service is done using JSON or YAML that follows the Swagger specification
schema20. A JSON Schema is available to validate specification documents21. The specification
defines all the standard parts of a RESTful API: the resource endpoints, supported functions (GET,
POST, DELETE, OPTIONS etc.), media types, parameters, and responses.
Swagger was used in the WaterML2.0 part 2 Interoperability Experiment to document the CSIRO
RESTful API22, as shown in Figure 15. The addition of inline parameter description and testing
provides an interactive document that is useful for developers. The service description is also
rendered as a RESTful API, which shows and example of the service description document23.
19 http://petstore.swagger.io/
20 http://swagger.io/specification/
21 https://github.com/swagger-api/validator-badge
22 http://waterml2.csiro.au/rgs-api/v1/docs
23
http://waterml2.csiro.au/rgs-api/v1/docs/api-docs/conversion

Figure 15 - Swagger UI of the WaterML2.0 part 2 RESTful API
5.1.2 RESTful API Modeling Language (RAML)
The RESTful API Modeling Language (RAML) is a non-proprietary open specification that is similar
to the Swagger specification in terms of its scope and implementation. It provides a JSON and
YAML description of RESTful APIs24.
5.1.3 RESTDesc
RESTDesc25 uses semantic approaches to define the functionality of hypermedia APIs. It uses the
RDF Turtle encoding to describe the resources and hypermedia links using existing semantic
vocabularies. In this sense it is simply a set of patterns for how to use existing vocabularies to
describe RESTful APIs.
5.1.4 Web Application Description Language (WADL)
The Web Application Description Language (WADL)26 is a W3C member submission (not a
standard), which provides an XML schema to describe Web API interfaces. It does this through
defining the API’s resources, the supported operations on each resource, and available query
24 http://raml.org/spec.html
25 http://restdesc.org/
26
http://www.w3.org/Submission/wadl/

parameters. The specification and underlying schema27 are not overly complicated and focus on
the core contracts for clients communicating with a Web API.
5.1.5 API Blueprint
Blueprint28 is an API description based on Markdown, which generates a JSON specification

format. The aim is to provide a more human friendly editing language (i.e. Markdown). Because it
is Markdown based, the specification is defined in the form of specific keywords combined with
section headers.
5.1.6 Mashery I/O Docs
Mashery29 has a suite of API offerings including services for API management, development,
documentation, security, and testing. The documentation engine is very similar to that of Swagger.
It offers generated HTML pages with forms to interact with an API. Its specification does not
appear to be open.
5.1.7 RESTful Service Description Language (RSDL)
The RESTful Service Description Language (RSDL)30 is an XML format for describing RESTful
services, including resources, media types, methods, links, authentication, parameters etc.
5.2 Specifications for API responses

The previous section covers how to describe a specific RESTful API, while this section covers
emerging approaches to structure Web API responses in a common way. The service description
approach would be used to describe an existing API; this section covers practices that may be
followed when designing an API from scratch. Generally the purpose of these specifications is to
support the REST concept of hypermedia as the engine of application state (HATEOAS).
The general idea behind HATEOAS is that any client should be able to navigate through available
application states by traversing links provided by the server. These links are dynamically provided
by the server and should not be hard-coded into clients. Thereby the server can drive application
behaviour without change to the client.
5.2.1 I-JSON
Internet JSON (I-JSON) is a specification that adds additional constraints to the IETF JSON31 to aid
interoperability and software compatibility for JSON on the Web. It is not a specific definition of
27 http://www.w3.org/Submission/wadl/wadl.xsd
28 https://apiblueprint.org/
29 http://www.mashery.com/api/io-docs
30 http://www.balisage.net/Proceedings/vol10/html/Robie01/BalisageVol10-Robie01.html
31
https://www.ietf.org/rfc/rfc4627.txt

response formatting, but a JSON subset that reduces variation and may be used by all Web APIs
that use JSON encodings on the Web.
The main requirements introduced in iJSON include:
 Encoding: I-JSON Messages must be encoded using UTF-8.
 Floating point numbers: numbers must be exactly representable as IEEE 754:2008 64-bit
binary floating point numbers
 No duplicate property names: JSON objects should not have properties with duplicate
names.
 Strict compliance: software receiving a JSON document that does not meet the iJSON
requirements should not trust or act on any content of the message.
There are also some recommendations that include: date time strings using ISO8601, binary data
encoded using base64, and use of objects or arrays as top-level constructs.
The OGC is considering recommending I-JSON for use within JSON encoded OGC standards.
5.2.2 Hypertext Application Language (HAL)
HAL32 provides a standardised response structuring for JSON and XML responses for Web APIs. The
basis for the structure is through the separation of resources, links, and embedded resources, as
shown in Figure 16.
32 http://stateless.co/hal_specification.html

Figure 16 - HAL structuring (source http://stateless.co/hal_specification.html)
HAL makes strong use of the hypermedia aspect of RESTful APIs through the use of link encodings.
It also supports URL templates33 in its link definitions. An example response is shown below.
{
"_links": {
"self": { "href": "/orders" },
"curies": [{ "name": "ea", "href": "http://example.com/docs/rels/{rel}", "templated": true }],
"next": { "href": "/orders?page=2" },
"ea:find": {
"href": "/orders{?id}",
"templated": true
},
"ea:admin": [{
"href": "/admins/2",
"title": "Fred"
}, {
"href": "/admins/5",
"title": "Kate"
}]
},
"currentlyProcessing": 14,
"shippedToday": 20,
"_embedded": {
"ea:order": [{
"_links": {
"self": { "href": "/orders/123" },
"ea:basket": { "href": "/baskets/98712" },
"ea:customer": { "href": "/customers/7809" }
},
"total": 30.00,
"currency": "USD",
"status": "shipped"
}, {
"_links": {
"self": { "href": "/orders/124" },
33
https://tools.ietf.org/html/rfc6570

"ea:basket": { "href": "/baskets/97213" },
"ea:customer": { "href": "/customers/12369" }
},
"total": 20.00,
"currency": "USD",
"status": "processing"
}]
}
}
HAL has been drafted as an IETF standard34.
5.2.3 Collection + JSON
Collection + JSON35 specifies a common structuring for retrieving, updating and deleting
collections of resources through an API. It does this through specifying patterns for using HTTP
Link Headers and providing a common structuring of collections with embedded objects. It has
similar goals to that of HAL.
A collection + JSON media type (application/vnd.collection+json) has been registered with IANA as
a common media type36. However the actual structuring has no status as a standard.
5.2.4 SIREN
SIREN37 is a structuring specification for JSON responses in Web APIs. Its scope is very similar to
that of HAL and Collection + JSON. It separates three key aspects: entities (resources), links and
actions. There is structure for embedding entities in the same way HAL does.
An example response is shown below.
{
"class": [ "order" ],
"properties": {
"orderNumber": 42,
"itemCount": 3,
"status": "pending"
},
"entities": [
{
"class": [ "items", "collection" ],
"rel": [ "http://x.io/rels/order-items" ],
"href": "http://api.x.io/orders/42/items"
},
{
"class": [ "info", "customer" ],
"rel": [ "http://x.io/rels/customer" ],
"properties": {
"customerId": "pj123",
"name": "Peter Joseph"
},
"links": [
{ "rel": [ "self" ], "href": "http://api.x.io/customers/pj123" }
]
34 https://tools.ietf.org/html/draft-kelly-json-hal-07
35 http://amundsen.com/media-types/collection/format/
36 http://www.iana.org/assignments/media-types/application/vnd.collection+json
37
https://github.com/kevinswiber/siren/blob/master/README.md

}
],
"actions": [
{
"name": "add-item",
"title": "Add Item",
"method": "POST",
"href": "http://api.x.io/orders/42/items",
"type": "application/x-www-form-urlencoded",
"fields": [
{ "name": "orderNumber", "type": "hidden", "value": "42" },
{ "name": "productCode", "type": "text" },
{ "name": "quantity", "type": "number" }
]
}
],
"links": [
{ "rel": [ "self" ], "href": "http://api.x.io/orders/42" },
{ "rel": [ "previous" ], "href": "http://api.x.io/orders/41" },
{ "rel": [ "next" ], "href": "http://api.x.io/orders/43" }
]
}
5.2.5 Uniform Basis for Exchanging Representations (UBER)
UBER38 is a document format that supports state transfers (hypermedia) and object encoding. It is
designed to be protocol-agnostic, with guidance on how to implement it using HTTP with XML or
JSON. It achieves this through use of reserved string properties that carry specific semantics in
terms of state transitions (actions), templates, and embedded or linked resources.
5.2.6 JSON API
JSON API39 is a specification of conventions for structuring JSON responses from Web APIs. It
addresses encoding issues such as content negotiation, document structure (links, compound
documents, metadata), sorting, pagination, and creating/updating/deleting resources. The
implementations appear to growing40 and the Github repository is very active. It addresses a
number of common issues across APIs, some of which have been identified in OGC REST activities.
5.2.7 Hydra
Hydra41 uses JSON-LD with a vocabulary42 to allow a server to advertise its state transitions to
clients, again to enable HATEOAS properties of Web APIs. The vocabulary defines the common
concepts within a RESTful API: Resources, Operations, Status Codes, Paged Collections, and Links.
A summary of the vocabulary is shown in Figure 17.
38 https://rawgit.com/uber-hypermedia/specification/master/uber-hypermedia.html
39 http://jsonapi.org/
40 http://jsonapi.org/implementations/
41 http://www.markus-lanthaler.com/hydra/
42
http://www.hydra-cg.com/spec/latest/core/

Figure 17 - Hydra core vocabulary for describing Web APIs
This vocabulary is then used to embed JSON-LD descriptions in JSON documents that describe a
Web API. There is an active W3C Community Group43 supporting the Hydra specification. The
Hydra vocabulary is used within the Linked Data Fragments44 implementations, which is a
generalisation on access mechanisms for Linked Data.
5.3 API management

API management is a broad term used to describe a number of features required for operating and
managing APIs through their lifecycle. These include the following types of functions:
 Analytics: report on usage, failed and slow requests, common queries etc.
 API portals: provides harmonised developer documentation, login, key management etc.
 Traffic and performance management: set request throttling, rate limits, quotas etc.
43 https://www.w3.org/community/hydra/
44
http://www.hydra-cg.com/spec/latest/linked-data-fragments/

 Service catalogs: discovery of services, status reporting.
 Security: authorisation and authentication, OAuth, single sign on, access control, key
creation etc.
 Deployment processes: version management, test/sandbox environments, deprecation,
etc.
These features are becoming increasingly important as organisations deploy more APIs, and look
to manage and monetise their APIs. There are many API management frameworks available, with
products focussing on specific features or supporting the suite of API management tools. An in
depth review of these is out of scope for this report, however a summary is provided in Table 6.
The table represents a very superficial review: no installations or testing of features was
performed. There are components that focus on particular features and some that have richer
support within a feature category. For example, ‘security’ can range from basic HTTP
authentication to OAuth2.0, single sign-on, and encryption etc. The categories are only to indicate
some level of support. A more in depth analysis would be required to make a well-informed
decision.

Table 6 - Basic comparison of API manager frameworks
Name Analytics API Traffic Service Security Deployment License Comments

portal Catalog
Tyk       Open Source,

cloud or
supported
WSO2 API     45   Open Source Part of the WSO2

Manager suite of Web
Services.
Gluu × × × ×  × Open Source

(multiple
packages)
API Axle  ×  ×  × Open Source Focussed on key

management.
Kong 46 ×47  48  ×49
45 With other WSO2 products

46 Via integration with Galileo
47 See https://github.com/Mashape/kong/issues/391, also available via integration with https://gelato.io/.
48 Via https://market.mashape.com/

3Scale API    ×  ? Closed source,
Management cloud-hosted
freemium
model
apigee Volos × ×  ×  × Open Source Supports cache

management
API Umbrella    ×  × Open Source
Repose  ×  ×  × Open Source Analytics is logging

only
StrongLoop (uses       Commercial Focused on creation

Loopback) with and deployment of
community NodeJS APIs.
(very limited)
edition.
API Man       Open Source Development

roadmap available
here
49 Via https://market.mashape.com

6 Conclusions
Web APIs hold great potential for agencies to provide on-demand, tailored and robust
environmental data. Accordingly, the design and implementation of Web APIs should be done with
careful consideration of existing practices. It is beneficial to be consistent with existing practices,
while being wary not to adopt practices that gain no traction.
It is obvious that most organisations are tackling the API design challenges. Many services have
evolved from the ground up and organisations are attempting to consolidate into more simple,
cohesive Web APIs. All these APIs have made simplifications and trade-offs when designing their
APIs to suit their intended audience. This report summarises a broad range of existing practices
with the aim of informing the design of future environmental data Web APIs. To complement the
analysis within this report, we provide some general observations and learnings based on our
experiences using and developing Web APIs.
6.1 General
Recommendation 1 – focus on core requirements
When designing APIs it is important to understand your users, while keeping in mind they may not
yet exist. Try to identify core requirements as a starting point, and build out as your user’s
requirements and needs grow.
Recommendation 2 – clearly separate functionality into separate APIs
Highly abstracted APIs can be difficult to understand and may be a barrier for new users. Trying to
do too many things in a single API or Web Service can result in an overly complex, hard to
understand API. Partitioning APIs into concrete API endpoints, each having clearly stated functions
help to build understanding.
RESTful APIs have a natural separation into ‘sub APIs’ through the use of the resource end points.
This can be a useful way of separating clear logical parts of the API that reflect nouns. For example
the GitHub API resources use very concrete nouns that reflect the nature of the API interaction –
see Figure 18.

Figure 18 - Example GitHub API resources
Recommendation 3 – productise your Web APIs

It appears to be beneficial to give Web APIs a product feel. They should be branded consistently
and behave consistently where there are multiple APIs, or endpoints provided.
Most of the popular Web APIs provide an initial documentation block that defines all the common
behaviour across all its resources/APIs. For example Figure 19 shows Stripe and GitHub’s base
topics, which cover many of the common API components described in this report.

Figure 19 - Stripe API (left) GitHub API (right)
Often successful APIs reflect the product of the company very closely. It is even sometimes the
case that a company’s product uses its own API, which can be very beneficial.
Recommendation 4 - publish targeted documentation
Informal, tutorial style, documentation appeals to developers and is the most commonly read
documentation. Avoid publishing only example URLs.
There are some excellent documentation engines available on the web. These provide navigation
structures, with narrative sections, and example code sections. For example, the Stripe API (a web-
based payment system) structures its documentation page into three parts: resources, description
and code examples; as shown in Figure 20.
Figure 20 - Stripe API documentation

This type of documentation gives developers everything they need to interact with an API and
rapidly get feedback on their interactions.
Many Web APIs have at a minimum the following core sections within their documentation:
 Quickstart: shortest possible path to getting something done. Developers thrive on getting
some quick results. Make it easy.
 Tutorials: longer walkthroughs that tackle specific goals, such as ‘discovering the temporal
range of available observations for a site’.
 API reference: all the gory details of your API. Often will be supported by a tool such as
Swagger or other doc-generators.
SDKS (software development kits): Links to client implementation of the API in different languages
or different platforms (e.g. mobile). Typically these point to Github pages with supporting tools.
Often consist of combination of organisation-developed and community-developed clients.
Recommendation 5 – foster a community
Provide community resources, such as forums and easy feedback mechanisms, to allow users and
developers to share experiences and issues. APIs are successful when they are used. Having an
organisational representative that is tasked with engaging with the community can be hugely
valuable.
Recommendation 6 – use domain abstractions and document them
Identify key simplifications and domain abstractions to reduce complexity. Clearly document these
to the users. This will avoid confusion.
Recommendation 7 – include your Web APIs in your enterprise architecture
Including your Web APIs in your enterprise architecture appears to be a practice that can bring
many benefits. They should not be ‘bolted on’, but rather may become a point of convergence for
multiple systems: a platform.
Recommendation 8 – document using common use cases
Document common usages of your APIs that achieve certain goals, especially when they require
multi-step requests to the API. For example, ‘I want to discover all observations available at a
particular station’.
6.2 Technical
Recommendation 9 – Use API documentation frameworks
Make use of automated API documentation where possible. These can often be synched directly
with an implementation version, which helps to minimise divergence. Some also provide
interactive (e.g. Swagger) documentation that allows inline requests to be made. This helps to
lower the barrier of entry for developers and quickly builds understanding.
Recommendation 10 – be cautious when selecting API description languages
Avoid cutting edge/emerging API description languages and response patterns. As discussed in 5.1
and 5.2, there has been an explosion in the number of these. Picking a winner is difficult. Using an
overly complex, non-supported service description framework and/or response structure can be
an impediment to developers. Ensure there is at least some uptake and solid usage before
adopting a practice.
Recommendation 11 – Consider using an API manager
There is a range of these available. Some may require customisation, or even coupling with other
technologies. A more in depth analysis of these is recommended if selecting one for use; along
with alignment of organisational IT practices (e.g. supported OS, languages, open source licencing
etc.). Generally they offer solutions to some very common problems in deployment and
management of APIs.
Recommendation 12 – Consider conforming to the I-JSON specification
The I-JSON Message Format (Bray, 2016) defines some precise restrictions (see 5.2.1) on the use of
JSON to aid interoperability, without adding implementation burden.
Recommendation 13 – Don’t do anything strange
There are many Web API best practices (see references), many of which agree on some common
patterns, some of which are already part of the HTTP standard. Always strive to use HTTP
consistently. The use of best practices requires consideration in each case. HTTP related practices
include:
 Structuring URIs (Masse, 2011a): use of slashes/hashes, prefer all lowercase, hyphens not
underscores, use of sub-domains.
 Use of HTTP request methods: use of GET, POST, PUT, DELETE and OPTIONS. Some
convergence is occurring on the use of these within Web APIs (Masse, 2011b).
 Use HTTP response status codes: these are generally well documented and their proper
use appears to be growing.
 Use HTTP headers, including (but not limited to):
o Use content-types and media-type syntax for specifying message content type,
o Use last-modified, cache-control, expires and date headers for proper cache
interactions,
o Use Location headers for URIs of newly created resources.
Recommendation 14 – Use consistent link representation and error messages
These are two areas where there is no widely agreed approach (see section 5.2). Even so, it is
important to choose an approach, clearly document it and be consistent across all Web APIs being
implemented.
Recommendation 15 – Define your versioning practices
There’s an extensive online debate (Dierking, 2016)(Zazueta, 2015)(Overflow, 2016) (Sahni, 2016)
about versioning RESTful APIs. The two prevailing options are to include a version in the URI (the
‘pragmatic approach’) or use HTTP headers (the ‘academic’ approach). Understand the trade-offs,
choose one, and stick with it.

6.3 Modelling the domain
Recommendation 16 – use concrete resource names
Using concrete concepts for resources with clear definitions aids understanding of APIs. Avoid
vague, internal concepts, which only have meaning in certain sub-domains.
Recommendation 17 – use design patterns for modelling the domain
Consider making use of domain-driven design patterns (Evans, 2003) when tackling domain
complexity. Designing Web APIs will inevitably involve interaction with many existing systems.
Patterns such as Anticorruption Layer, Ubiquitous Language, Domain Vision Statement and
Bounded Context (Evans, 2003) all provide useful starting points for solving domain modelling and
integration problems.
Recommendation 18 – involve the experts
Involve domain specialists in resource design. In organisations as complex as multi-disciplinary
environmental agencies, there won’t be a single information model or Web API to rule them all.
Rather there exists multiple domain contexts (Bounded Contexts - (Fowler, 2014) and (Evans,
2003)) and words with different meaning (polysemes), e.g. what is a monitoring station? Any
domain-level simplifications should be reviewed within these contexts to ensure they make sense.
Recommendation 19 – use abstractions for common sampling regimes
Consider point, grid and trajectory abstractions. This has been used effectively in existing APIs (e.g.
DAP, NetCDF) as a way of delineating different data types.
Recommendation 20 – Seek help from active groups
Monitor the outcomes of the joint W3C/OGC ‘Spatial Data on the Web’ working group50. It is
working on general recommendations for publishing spatial data on the web in a consistent
manner. These should be very useful in guiding convergence in some of the issues covered in this
report.
50
http://www.w3.org/2015/spatial/wiki/Main_Page

Appendix A
Summary of example data products from the Australian Bureau of Meteorology

id title url Blurb active Type of service
Australia rainfall The rainfall and river conditions website provides rainfall and
and river http://www.bom.gov river data collected for flood warning purposes from a network
7 conditions .au/australia/flood/ of stations across Australia. Y Web application;
http://www.bom.gov This project is designed to monitor sea level around the
Australian Baseline .au/oceanography/pr coastline of Australia, with an emphasis on the enhanced
Sea Level ojects/abslmp/abslm greenhouse effect. Selected hourly, monthly and yearly reports PDF reports; CSV
8 Monitoring Project p.shtml are currently available. Y download
http://www.bom.gov
.au/watl/about-
weather-and-
climate/australian- Information;
Australian climate climate- This online schematic diagram presents the main influences on educational
13 influences influences.shtml the Australian climate. Y content.
Australian Climate
Observations
Reference
Network - Surface
Air Temperature http://www.bom.gov PDF reports; CSV
(ACORN-SAT) .au/climate/change/a The ACORN-SAT dataset provides a record of daily download; web
14 dataset corn-sat/ temperatures over the last 100 years. Y application
Australian Geofabric is a specialised Geographic Information System. It Spatial data
Hydrological http://www.bom.gov registers the spatial relationships between important download via FTP;
Geospatial Fabric .au/water/geofabric/i hydrological features such as rivers, water bodies, aquifers and web application;
19 (Geofabric) ndex.shtml monitoring points. Y OGC web
services.
Australian
Meteorological
and The Australian Meteorological and Oceanographic Journal
Oceanographic http://www.bom.gov presents articles for the atmospheric, oceanic and related Journal;
21 Journal .au/amoj/index.shtml sciences on topics relevant to the Southern Hemisphere. Y information.
Australian
Network of
Hydrologic This interactive tool allows users to interrogate, display and Web application;
Reference Stations http://www.bom.gov download data about streamflow and streamflow variability in CSV download;
23 web portal .au/water/hrs/ Australia. Y analysis
Australian Water Australian Water Resources Assessment reports highlight
Resources http://www.bom.gov patterns, trends and variability in water quantity at regional to PDF reports; CSV
27 Assessments .au/water/awra/ national scales and over time from months to decades. Y download
Information;
Climate Change in Australia shows how Australia's climate may educational
change in the future. Climate change projections are available content; data
http://www.climatec for low, mid-range and high greenhouse gas scenarios - drawing download (CSV,
Australian climate hangeinaustralia.gov. on scenarios developed for the Intergovernmental Panel on JSON, NetCDF);
34 change projections au Climate Change (IPCC) Fourth Assessment Report. Y analysis
http://www.bom.gov Climate Data Online provides access to a range of statistics,
Climate Data .au/climate/data/ind historical weather observations, climatology maps, and other CSV download;
35 Online ex.shtml Australian climate data. Y PDF tables.
Information;
http://www.bom.gov The Climate Model Summary provides Pacific and Indian Ocean educational;
Climate Model .au/climate/ahead/m temperature outlooks for the next six months based on a survey model
36 Summary odel-summary.shtml of international climate models. Y exploration.
Climate Maps (images);
statements - http://www.bom.gov Climate statements provide links to monthly, seasonal and data summaries
monthly, seasonal, .au/climate/current/i annual climate summaries routinely issued by the Bureau of (tables); CSV
37 annual ndex.shtml Meteorology. Y summaries; PDFs
42 Drought http://www.bom.gov The Drought Statement provides a monthly summary of rainfall Y Maps (images).

Statement .au/climate/drought/ deficiencies in Australia.
drought.shtml
El Niño-
Southern Information
Oscillation Wrap- http://www.bom.gov The ENSO Wrap-Up provides information on El Niño/La (text); graphs
46 Up .au/climate/enso/ Niña weather patterns for Australia. Y (images);
http://www.bom.gov enGauge is an online newsletter about all things associated with
enGauge .au/water/about/pub water, such as dam water levels, new policies and regulations
49 newsletter lications/index.shtml and catchment information. Y Newsletters
http://www.bom.gov
Fire weather .au/weather- Fire weather warnings are issued when the weather conditions Warnings; RSS,
54 warnings services/bushfire/ are conducive to the spread of dangerous bushfires. Y XML, AMOC
Forecast Explorer provides an interactive web display of
Forecast Explorer http://www.bom.gov forecasts out to seven days for Victoria, Tasmania, South
(RETIRED AT 14 .au/australia/index.s Australia and New South Wales. Other states and territories will
55 AUGUST) html be added 2013-2015. N Retired.
http://www.bom.gov Frost potential maps show forecast low temperature thresholds
.au/jsp/watl/weather for various Australian weather station locations. This is a trial
57 Frost potential /frost.jsp service. Y Web application;
http://www.bom.gov High seas forecasts are issued twice daily and include
.au/marine/high- information on wind, waves, temperature and rain for areas Web application;
60 High seas forecasts seas.shtml beyond the coastal waters surrounding Australia. Y text data.
The Joint Australian Tsunami Warning Centre is operated 24
Joint Australian http://www.bom.gov hours a day and detects, monitors, verifies and warns of any
Tsunami Warning .au/tsunami/index.sh tsunami threat to the coastline of Australia and its offshore
62 Centre tml territories. Y Web application;
Madden-Julian Madden-Julian Oscillation (MJO) monitoring provides data on
Oscillation http://www.bom.gov major fluctuations in tropical weather, attributed to this Web application;
67 monitoring .au/climate/mjo/ weather pattern. Y images; PDF
http://www.bom.gov Marine wind forecasts are issued twice daily and provide
Marine wind .au/marine/wind.sht graphical information on wind direction and speed for different
68 forecast ml marine areas across Australia. Y Web application;

A series of documents provide monthly overviews of
Monthly Weather http://www.bom.gov temperatures, rainfall and significant weather events in
72 Review .au/climate/mwr/ Australia. Y PDF reports
National Atlas of The National Atlas of Groundwater Dependent Ecosystems
Groundwater http://www.bom.gov displays ecological and hydrogeological information of known
Dependent .au/water/groundwa groundwater dependent ecosystems and ecosystems that
75 Ecosystems ter/gde/index.shtml potentially use groundwater. Y Web application;
Web application;
National seasonal http://www.bom.gov National seasonal temperature outlooks are general statements spatial data
temperature .au/climate/ahead/te about the probability or risk of wetter or drier than average download (shape
82 outlook mps_ahead.shtml weather over a three-month period. Y file, KMZ)
The National Water Account provides standardised annual
information about the management and use of Australia's
National Water http://www.bom.gov water resources, including water stores and flows, water rights
85 Account .au/water/nwa/ and water use. Y PDF reports
Ocean forecasts: http://www.bom.gov
sea temperatures .au/oceanography/fo Modelled ocean forecasts provide information on sea
88 and currents recasts/ temperatures and currents across Australian waters. Y Web application;
http://www.bom.gov
.au/climate/pacific/a The Pacific Climate Change Data Portal provides access to raw
Pacific Climate bout-pacific-data- and homogenised temperature and rainfall data for several Web application;
90 Change Data Portal portal.shtml countries and islands in the South Pacific region. Y CSV download;
Predictive Ocean
Atmosphere The Predictive Ocean Atmosphere Model for Australia (POAMA) Seasonal model;
Model for http://poama.bom.g is a seasonal to inter-annual seasonal forecast modelling system NetCDF/THREDDS
91 Australia ov.au/ that provides a number of experimental operational products. Y access.
Rainfall and http://www.bom.gov Rainfall and temperature records provide the highest and
temperature .au/climate/extreme/ lowest temperatures and rainfall levels for each Australian state HTML tables; PDF
93 records records.shtml and territory. Y reports
http://www.bom.gov Rainfall forecasts, over a five-day period, are produced through
.au/jsp/watl/rainfall/ computer modelling and published graphically, for districts
94 Rainfall forecast pme.jsp around Australia. Y Web application;

Seasonal climatic outlooks are produced on a three-monthly
basis and provide information on rainfall, temperature, El Information;
Seasonal climatic http://www.bom.gov Niño/La Niña status, streamflow, tropical cyclones analysis; links to
97 outlooks .au/climate/ahead/ and climate models. Y web applications;
Seasonal http://www.bom.gov Streamflow forecasts are presented graphically for north- Web application;
streamflow .au/water/ssf/index.s eastern Australia, providing the likelihood of low, near median CSV download;
98 forecasts html or high streamflows over a three-month period. Y analysis
http://www.bom.gov The Southern Oscillation Index indicates the development and
Southern .au/climate/current/s intensity of El Niño or La Niña events in the Pacific Graphs (images);
102 Oscillation Index oi2.shtml Ocean. Y CSV download
http://www.bom.gov
Spatial sea surface .au/oceanography/oc Spatial sea surface temperature anomaly forecasts are plotted
temperature eantemp/GBR_SST.sh for the Great Barrier Reef using the Bureau of Meteorology's Graphs (images);
103 anomaly forecast tml seasonal prediction system. Y information
http://www.bom.gov Special climate statements provide a detailed summary of
Special climate .au/climate/current/s significant (and unusual) weather and climatic events in
104 statements tatements/ Australia. Y PDF reports
Tide predictions
for Australia, This tide predictions website provides estimated low and high
South Pacific and http://www.bom.gov tide levels and times for ports within Australia, the South Pacific Web application;
110 Antarctica .au/australia/tides/ and Antarctica. Y PDF reports
http://www.bom.gov
Tropical Cyclone .au/nt/forecasts/tcou The Tropical Cyclone Outlook is a regular report on existing and Not running
112 Outlook tlook.shtml potential cyclones in Australia. Y (summer only)
The National UV Forecast Graph shows the maximum clear sky
http://www.bom.gov UV Index at noon for the whole of Australia. The State UV
Ultraviolet index .au/australia/uv/inde Forecast Map shows UV Index forecasts for specific locations in Web application;
114 forecast x.shtml states and territories. Y images
The water storage website (and associated iPhone app) allows
http://water.bom.go users to compare water storage levels and volumes for 250
v.au/waterstorage/a publically-owned lakes, reservoirs and weirs in states and
118 Water storage wris/ territories and to see how much water is available in Australia. Y Web application;

http://www.bom.gov The analysis chart archive provides access to weather maps for
Weather analysis .au/australia/charts/ the Australian, Southeast Asian/Western Pacific and Southern
120 chart archive archive/index.shtml Hemisphere regions. Y Images
Weather forecasts, http://www.bom.gov These services encompass a wide range of forecasts, warnings,
warnings and .au/australia/index.s weather observations, radar images and information services
121 observations html for the Australian continent, Antarctica and the South Ocean. Y Landing page.
http://www.bom.gov
Weekly rainfall .au/climate/current/ The weekly rainfall update provides a detailed analysis and map Information;
123 update weeklyrain.shtml of rainfall recorded across Australia during the past week. Y images
http://www.bom.gov The Weekly Tropical Climate Note is a summary report and
Weekly Tropical .au/climate/tropnote forecast of Australia's tropical climate, produced on a weekly
124 Climate Note /tropnote.shtml basis. Y Information (text)
Web application;
The Australian Water Availability Project (AWAP) monitors the images;
state and trend of the terrestrial water balance of the NetCDF/Thredds
Australian Water http://www.eoc.csiro Australian continent. The AWAP website provides water to registered
129 Availability Project .au/awap/ balance maps, reports and data files. Y users.
MetEye is an online mapping tool used to visualise weather
http://www.bom.gov observations and forecasts for Australia. It presents information
.au/australia/meteye on temperature, humidity, rainfall, wind, sea surface
132 MetEye / temperature, river conditions and cloud cover. Y Web application;
This website provides links to information about weather
Climate and past conditions in Australia–past, future and present. It
weather includes links to seasonal outlooks, reports, summaries and
information and http://www.bom.gov maps. Data services provide past weather and climate
138 data services .au/climate/ information from the Bureau's Australian climate data archive. Y Landing page.
http://www.bom.gov The Severe Storms Archive contains data relating to recorded
Severe Storms .au/australia/stormar severe thunderstorm and related events in Australia dating Web application;
142 Archive chive/ back to the 18th Century. Y CSV download;
http://www.bom.gov The Marine Water Quality Dashboard is a tool to access and Web application;
Marine Water .au/marinewaterqual visualise a range of water quality indicators for the Great CSV, NetCDF,
145 Quality Dashboard ity/ Barrier Reef. It enables access to near real-time data, and over Y KMZ, GeoTIFF

ten years of information, on sea surface temperatures,
chlorophyll levels, suspended sediments, and coloured
dissolved organic matter.
eXchange is an electronic newsletter about environmental
http://www.bom.gov information activity with a focus on Australian Government
.au/environment/pu developments, including strategic activities, new and existing
152 eXchange blications.shtml products and services and major events. Y Newsletters
Information (text,
This website provides information on the science of climate images);
Climate change http://www.bom.gov change and climate variability. It includes climate updates, educational;
153 and variability .au/climate/change/ trend maps and information about relevant datasets. Y analysis
Meteorological analyses for Australia were generated as a
contribution to the Australian Water Availability Project. They
Climate maps - include rainfall, temperature, vapour pressure, solar exposure Web application;
Australian Water http://www.bom.gov and the Normalised Difference Vegetation Index (NDVI) over images; PDF;
155 Availability Project .au/jsp/awap/ time periods ranging from daily, weekly, monthly to 3-yearly. Y GRID download.
http://www.bom.gov This mobile phone application provides weather, temperature,
.au/uv/iphoneapp.sh UV level and sun protection times for the day anywhere in
156 SunSmart app tml Australia. Y Mobile app.
Indigenous The indigenous weather knowledge website displays seasonal Information
weather http://www.bom.gov weather calendars, developed over thousands of years by (text);
157 knowledge .au/iwk/ Indigenous communities. Y educational;
Radio and space http://www.ips.gov.a This website provides the latest space weather observations
159 weather services u/ and forecasts. Users can also subscribe to reports and alerts. Y Text, images,
Weather http://bom- This website provide an online weather community where
Observations wow.metoffice.gov.u Australians can share weather observations, information and Web application;
160 Website (WOW) k/ photos. Y community data.
http://www.bom.gov The Bureau of Meteorology operates a nationwide network of
Weather Watch .au/australia/radar/? over 60 Weather Watch radars. This website provides radar Web application;
161 radars ref=ftr locations, images, news and education. Y images
http://www.bom.gov The water restrictions website provides information about Web application;
163 Water restrictions .au/water/restriction current water restrictions around Australia. It is searchable by Y text data.

s/ State or Territory, water agency and restriction name.
http://www.bom.gov This website provides information about the Bureau of
Weather Station .au/climate/data/stat Meteorology weather stations, which can help users to select Web application;
164 Directory ions/ the type of data they require. Y CSV download;
The Bureau of Meteorology flood forecasting and warning
National flood service uses rainfall and streamflow observations, numerical
forecasting and http://www.bom.gov weather predictions and hydrologic models to forecast and
165 warning service .au/water/floods/ warn for possible flood events across Australia. Y Information; PDFs
http://www.bom.gov These maps show the variability of the Australian climate by
Maps of average .au/climate/averages providing information about average as well as more extreme
166 conditions /maps.shtml conditions. Y Maps (images).
http://www.bom.gov
Current tropical .au/cyclone/?ref=dro This website provides information and warnings about current Information;
167 cyclones pdown tropical cyclones in Australia and its waters. Y images (maps)
Pilot heatwave http://www.bom.gov This website provides a graphical map of heatwaves, severe
forecast for .au/australia/heatwa heatwaves and extreme heatwaves for the current day
168 Australia ve/ extending out for the next four days. Y N/A
Summary of Web APIs

Hype
rmed Interface description
Name Encodings ia HTTP verbs Data types language URL
NOAA Climate https://www.ncdc.noaa.gov/cdo-
Data API v2 JSON No GET Observations, stations Text-based web/webservices/v2
UK Met Office Observations, forecasts,
- DataPoint JSON, XML, Images No GET stations, imagery, text Text-based http://www.metoffice.gov.uk/datapoint
Observations,
Atlas of Living GET, POST, species/taxon data,
Australia JSON Yes DELETE spatial data, Dynamic HTML http://api.ala.org.au/
UK Bathing Linked Data* Yes GET Obserations, spatial Dynamic linked-data http://www.epimorphics.com/web/wiki/

water quality (zones, sampling points), HTML bathing-water-quality-structure-
assessments published-linked-data
UK National
Biodiversity Unkn GET, PUT, https://data.nbn.org.uk/Documentation/
Network JSON, XML own DELETE Observations, taxon WADL Web_Services/Web_Services-REST/
GeoServices JSON schema +
API JSON No GET, POST Spatial, images specification http://portal.opengeospatial.org
Global Change
Information Observations, model
System JSON Yes GET outputs, reports, tables.. Swagger (dynamic HTML) http://data.globalchange.gov/api/
GET, POST,
JSON (API) + PUT, DELETE,
CKAN various Yes PATCH Range (generic) Text-based http://docs.ckan.org/en/latest/
Data Access
Protocol / HTTP Text + binary Som Spatial (grids), http://www.opendap.org/pdf/ESE-RFC-
Hyrax (octet) e GET observations (tables) Text-based (specification) 004v1.2.pdf
THREDDS Multiple services
(multiple Spatial (grids), (standards: WMS, WCS, http://www.unidata.ucar.edu/software/t
services) XML, NetCDf Yes GET observations (tables) NC subsetting etc). hredds/current/tds/
ERDDAP JSON, XML,
(multiple NetCDF, CSV, Spatial (grids), http://coastwatch.pfeg.noaa.gov/erddap
services) matlab, xhtml Yes GET observations (tables) Text-based /index.html
Forecast.io JSON No GET Observations, forecasts Text-based https://developer.forecast.io/docs/v2

USGS Water
Data Services JSON, CSV, XML No GET Observations, statistics Text-based http://waterservices.usgs.gov/
HTML, CSV, JSON,

*Linked Data RDF, text, TTL,
encodings XHTML, XML

References
Bechhofer, S., J. Ainsworth, J. Bhagat, I. Buchan, P. Couch, D. Cruickshank, D. De Roure, et al. 2010.
“Why Linked Data Is Not Enough for Scientists.” In , 300–307. doi:10.1109/eScience.2010.21.
Bray, Tim. 2016. “The I-JSON Message Format.” https://tools.ietf.org/html/rfc7493.
Brooks, F. P. 1987. No Silver Bullet. April.
https://books.google.com.au/books?hl=en&lr=&id=eT7uBwAAQBAJ&oi=fnd&pg=PA11&dq=
no+silver+bullet&ots=bUfoIuR0Oz&sig=XGUpG1-_9A-XjdaG4X8tqDuoa4k.
Bureau of Meteorology. 2015. “Bureau of Meteorology Strategic Plan 2015-2020.”
http://www.bom.gov.au/info/leaflets/strategic-plan-2015-20.pdf.
Cartwright, John, Jesse Varner, and Susan McLean. 2015. “Data Stewardship: How NOAA Delivers
Environmental Information for Today and Tomorrow.” Marine Technology Society Journal 49
(2): 107–11.
Cornillon, P., J. Gallagher, and T. Sgouros. 2003. “OPeNDAP: Accessing Data in a Distributed,
Heterogeneous Environment.” Data Science Journal 2: 164–74. doi:10.2481/dsj.2.164.
Cox, Simon JD, Jonathan Yu, and Terry Rankine. 2014. “SISSVoc: A Linked Data API for Access to
SKOS Vocabularies.” Semantic Web 7 (1): 9–24.
Cox, Simon, and Peter Taylor. 2015. “Observations & Measurements JSON Encoding - Discussion
Paper. OGC 15-100r1.” 15-100.
D’Amore, F., S. Cinnirella, and N. Pirrone. 2013. “Data and Metadata Management Automation for
an Effective Approach to Sharing Environmental Data.” Edited by N. Pirrone. Proceedings of
the 16th International Conference on Heavy Metals in the Environment 1: 18003.
doi:10.1051/e3sconf/20130118003.
Dierking, Howard. 2016. “Versioning RESTful Services | Howard Dierking.”
http://codebetter.com/howarddierking/2012/11/09/versioning-restful-services/.
DiGiuseppe, Nicholas, Line C. Pouchard, and Natalya F. Noy. 2014. “SWEET Ontology Coverage for
Earth System Sciences.” Earth Science Informatics 7 (4): 249–64.
Dow, Andrew K., Eli M. Dow, Thomas D. Fitzsimmons, and Maurice M. Materise. 2015.
“HARNESSING THE ENVIRONMENTAL DATA FLOOD A Comparative Analysis of Hydrologic,
Oceanographic, and Meteorological Informatics Platforms.” Bulletin of the American
Meteorological Society 96 (5): 725–36. doi:10.1175/BAMS-D-13-00178.1.
Elkhatib, Y., G.S. Blair, and B. Surajbali. 2013. “Experiences of Using a Hybrid Cloud to Construct an
Environmental Virtual Observatory.” In , 13–18.
Evans, Eric. 2003a. “BOUNDED CONTEXT.” In Domain-Driven Design: Tackling Complexity in the
Heart of Software. Addison-Wesley Professional.
http://proquest.safaribooksonline.com/book/project-
management/0321125215/fourteendot-maintaining-model-integrity/ch14lev1sec1.
Evans, Eric. 2003. Domain-Driven Design: Tackling Complexity in the Heart of Software. Addison-
Wesley Professional. http://proquest.safaribooksonline.com/book/project-
management/0321125215.
Fowler, Marting. 2003. “MultipleCanonicalModels.” Martinfowler.com.
http://martinfowler.com/bliki/MultipleCanonicalModels.html.
Fowler, Marting. 2014. “BoundedContext.” Martinfowler.com.
http://martinfowler.com/bliki/BoundedContext.html.
Gong, Jianya, Jing Geng, and Zeqiang Chen. 2015. “Real-Time GIS Data Model and Sensor Web
Service Platform for Environmental Data Management.” International Journal of Health
Geographics 14 (1): 2. doi:10.1186/1476-072X-14-2.
Heritage, Laura. 18:07:26 UTC. “API Description Languages.”
http://www.slideshare.net/SOA_Software/api-description-languages.
Hilbring, Desiree, Anastasia Moumtzidou, Juergen Mossgraber, and Stefanos Vrochidis. 2014.
“Semantically Enriching an Open Source Sensor Observation Service Implementation for
Accessing Heterogeneous Environmental Data Sources.” Transactions in Gis 18 (4): 480–95.
doi:10.1111/tgis.12055.
H, Jeremy. 2011. “There Is No Right Way: Versioning and Types in REST/HTTP API Resources.”
There Is No Right Way. February 24.
http://thereisnorightway.blogspot.com.au/2011/02/versioning-and-types-in-resthttp-
api.html.
Horsburgh, Jeffery S., David G. Tarboton, Michael Piasecki, David R. Maidment, Ilya Zaslavsky,
David Valentine, and Thomas Whitenack. 2009. “An Integrated System for Publishing
Environmental Observations Data.” Environmental Modelling & Software 24 (8): 879–88.
doi:10.1016/j.envsoft.2009.01.002.
Houbie, Frédéric. 2015. “Testbed 11 REST Interface Engineering Report.” OGC Engineering Report
OGC 15-052.
Jacobson, Daniel, Greg Brail, and Dan Woods. 2011a. APIs: A Strategy Guide. O’Reilly Media, Inc.
http://proquest.safaribooksonline.com/book/information-technology-and-software-
development/9781449321628.
Jacobson, Daniel, Greg Brail, and Dan Woods. 2011. “Defining the Value Chain: Ask Key
Questions.” In APIs: A Strategy Guide. O’Reilly Media, Inc.
development/9781449321628/5dot-key-design-principles-for-apis/id2809367.
Jacobson, Daniel, Greg Brail, and Dan Woods. 2011. “Why You Might Need an API.” In APIs: A
Strategy Guide. O’Reilly Media, Inc.
development/9781449321628/2dot-apis-as-a-business-strategy/id2743825.

Jazayeri, Mohammad Ali, Steve H. L. Liang, and Chih-Yuan Huang. 2015. “Implementation and
Evaluation of Four Interoperable Open Standards for the Internet of Things.” Sensors 15 (9):
24343–73. doi:10.3390/s150924343.
Kelly, Mike. 2015. “JSON Hypertext Application Language.” Accessed December 1.
https://tools.ietf.org/html/draft-kelly-json-hal-07.
Klabnik, Steve. 2015. “UBER: A New Hypermedia Format for APIs.” InfoQ. Accessed December 1.
http://www.infoq.com/news/2014/05/uber-hypermedia.
Kotsev, Alexander, Francesco Pantisano, Sven Schade, and Simon Jirka. 2015. “Architecture of a
Service-Enabled Sensing Platform for the Environment.” Sensors 15 (2): 4470–95.
doi:10.3390/s150204470.
Lanthaler, Markus. 2013. “Creating 3rd Generation Web APIs with Hydra.” In Proceedings of the
22nd International Conference on World Wide Web Companion, 35–38. International World
Wide Web Conferences Steering Committee. http://dl.acm.org/citation.cfm?id=2487799.
Lanthaler, Markus. 2014. “Leveraging Linked Data to Build Hypermedia-Driven Web APIs.” In REST:
Advanced Research Topics and Practical Applications, 107–23. Springer.
http://link.springer.com/chapter/10.1007/978-1-4614-9299-3_7.
Lanthaler, Markus, and Christian Gütl. 2012. “On Using JSON-LD to Create Evolvable RESTful
Services.” In Proceedings of the Third International Workshop on RESTful Design, 25–32. WS-
REST ’12. New York, NY, USA: ACM. doi:10.1145/2307819.2307827.
Lanthaler, Markus, and Christian Gütl. 2013. “Hydra: A Vocabulary for Hypermedia-Driven Web
APIs.” LDOW 996.
http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.362.4758&rep=rep1&type=pdf.
Lanthaler, Markus, and Christian Gütl. 2013. “Model Your Application Domain, Not Your JSON
Structures.” In Proceedings of the 22nd International Conference on World Wide Web
Companion, 1415–20. International World Wide Web Conferences Steering Committee.
http://dl.acm.org/citation.cfm?id=2488184.
Lensmar, Ole. 2013. “An Overview of REST Metadata Formats.” API UX.
http://apiux.com/2013/04/09/rest-metadata-formats/.
Liang, Steve, Chih-Yuan Huang, and Tania Khalafbeigi. 2015. “OGC® SensorThings API - Part 1:
Sensing.” Candidate standard 15-078r4. OGC.
Masse, Mark. 2011a. “Identifier Design with URIs.” In REST API Design Rulebook. O’Reilly Media,
Inc. http://proquest.safaribooksonline.com/book/web-development/9781449317904/2dot-
identifier-design-with-uris/id2814545.
O’Reilly. 2011b. “Request Methods.” In REST API Design Rulebook. O’Reilly Media, Inc.
http://proquest.safaribooksonline.com/book/web-development/9781449317904/2dot-
O’Reilly. 2011c. “URI Authority Design.” In REST API Design Rulebook. O’Reilly Media, Inc.
http://proquest.safaribooksonline.com/book/web-development/9781449317904/2dot-

Ma, Xiaogang, Peter Fox, Curt Tilmes, Katharine Jacobs, and Anne Waple. 2014. “Capturing
Provenance of Global Change Information.” Nature Climate Change 4 (6): 409–13.
doi:10.1038/nclimate2141.
Michener, William K. 2015. “Ecological Data Sharing.” Ecological Informatics 29, Part 1
(September): 33–44. doi:10.1016/j.ecoinf.2015.06.010.
National Water Commission. 2006. “Australian Water Resources Information System: User
Requirements.”
Nativi, Stefano, Paolo Mazzetti, Mattia Santoro, Fabrizio Papeschi, Max Craglia, and Osamu Ochiai.
2015. “Big Data Challenges in Building the Global Earth Observation System of Systems.”
Environmental Modelling & Software 68 (June): 1–26. doi:10.1016/j.envsoft.2015.01.017.
Nottingham, Mark, Marc Hadley, Joe Gregorio, David Orchard, and Roy T. Fielding. 2015. “URI
Template.” Accessed December 1. https://tools.ietf.org/html/rfc6570.
Overflow, Stack. 2016. “Best Practices for API Versioning?”
http://stackoverflow.com/questions/389169/best-practices-for-api-versioning.
Overpeck, Jonathan T., Gerald A. Meehl, Sandrine Bony, and David R. Easterling. 2011. “Climate
Data Challenges in the 21st Century.” Science 331 (6018): 700–702.
doi:10.1126/science.1197869.
Pääkkönen, Pekka, and Daniel Pakkala. 2015. “Reference Architecture and Classification of
Technologies, Products and Services for Big Data Systems.” Big Data Research. Accessed
August 5. doi:10.1016/j.bdr.2015.01.001.
Pautasso, Cesare. 2014. “RESTful Web Services: Principles, Patterns, Emerging Technologies.” In
Web Services Foundations, edited by Athman Bouguettaya, Quan Z. Sheng, and Florian
Daniel, 31–51. Springer New York. http://link.springer.com/chapter/10.1007/978-1-4614-
7518-7_2.
Pautasso, Cesare, Olaf Zimmermann, and Frank Leymann. 2008. “Restful Web Services vs. Big’web
Services: Making the Right Architectural Decision.” In Proceedings of the 17th International
Conference on World Wide Web, 805–14. ACM. http://dl.acm.org/citation.cfm?id=1367606.
“PharkMillups/beautiful-Docs.” 2016a. GitHub. Accessed February 10.
https://github.com/PharkMillups/beautiful-docs.
Pouchard, Line C., Marcia L. Branstetter, Robert B. Cook, Ranjeet Devarakonda, Jim Green, Giri
Palanisamy, Paul Alexander, and Natalya F. Noy. 2013. “A Linked Science Investigation:
Enhancing Climate Change Data Discovery with Semantic Technologies.” Earth Science
Informatics 6 (3): 175–85.
Reis, Stefan, Edmund Seto, Amanda Northcross, Nigel W. T. Quinn, Matteo Convertino, Rod L.
Jones, Holger R. Maier, et al. 2015. “Integrating Modelling and Smart Sensors for
Environmental and Human Health.” Environmental Modelling & Software. Accessed July 21.
doi:10.1016/j.envsoft.2015.06.003.
Rettig, Andrew J., Sumit Khanna, and Richard A. Beck. 2015. “Open Source REST Services for
Environmental Sensor Networking.” Applied Geography 60 (June): 294–300.
doi:10.1016/j.apgeog.2014.11.003.

Romañach, Stephanie S., Mark McKelvy, Kevin Suir, and Craig Conzelmann. 2015. “EverVIEW: A
Visualization Platform for Hydrologic and Earth Science Gridded Data.” Computers &
Geosciences 76 (March): 88–95. doi:10.1016/j.cageo.2014.12.004.
Sahni, Vinay. 2016. “Best Practices for Designing a Pragmatic RESTful API.” Vinay Sahni.
http://www.vinaysahni.com/best-practices-for-a-pragmatic-restful-api.
Schaaf, Hylke van der, and Reinhard Herzog. 2015. “Mapping the OGC SensorThings API onto the
OpenIoT Middleware.” In Interoperability and Open-Source Solutions for the Internet of
Things, edited by Ivana Podnar Žarko, Krešimir Pripužić, and Martin Serrano, 62–70. Lecture
Notes in Computer Science 9001. Springer International Publishing.
Schandl, Thomas, and Andreas Blumauer. 2010. “PoolParty: SKOS Thesaurus Management Utilizing
Linked Data.” In The Semantic Web: Research and Applications, 421–25. Springer.
http://link.springer.com/10.1007%2F978-3-642-13489-0_36.
Swain, Nathan R., Kilisimasi Latu, Scott D. Christensen, Norman L. Jones, E. James Nelson, Daniel P.
Ames, and Gustavious P. Williams. 2015. “A Review of Open Source Software Solutions for
Developing Water Resources Web Applications.” Environmental Modelling & Software 67
(May): 108–17. doi:10.1016/j.envsoft.2015.01.014.
Taylor, Peter, Paul Sheahan, David Briar, Stuart Hamilton, Brian Gouge, Peter Heweston, and
Matthew Fry. 2015. “WaterML2.0 Part 2 - RESTful API and JSON Encoding.” OGC Discussion
Paper 15-033. https://portal.opengeospatial.org/files/?artifact_id=63302&version=1.
Usländer, Thomas. 2005. “Trends of Environmental Information Systems in the Context of the
European Water Framework Directive.” Environmental Modelling & Software, Environmental
Knowledge and Information SystemsEnvironmental Knowledge and Information Systems, 20
(12): 1532–42. doi:10.1016/j.envsoft.2004.09.029.
Usländer, Thomas, Patrick Jacques, Ingo Simonis, and Kym Watson. 2010. “Designing
Environmental Software Applications Based upon an Open Sensor Service Architecture.”
Environmental Modelling & Software, Thematic issue on Sensors and the Environment –
Modelling & ICT challenges, 25 (9): 977–87. doi:10.1016/j.envsoft.2010.03.013.
Verborgh, Ruben, Thomas Steiner, Rik Van de Walle, and Joaquim Gabarro. 2012. “Linked Data and
Linked APIs: Similarities, Differences, and Challenges.” In The Semantic Web: ESWC 2012
Satellite Events, edited by Elena Simperl, Barry Norton, Dunja Mladenic, Emanuele Della
Valle, Irini Fundulaki, Alexandre Passant, and Raphaël Troncy, 272–84. Lecture Notes in
Computer Science 7540. Springer Berlin Heidelberg.
Vitolo, C., Y. Elkhatib, D. Reusser, C.J.A. Macleod, and W. Buytaert. 2015. “Web Technologies for
Environmental Big Data.” Environmental Modelling and Software 63: 185–98.
doi:10.1016/j.envsoft.2014.10.007.
Wanner, Leo, Harald Bosch, Nadjet Bouayad-Agha, Gerard Casamayor, Thomas Ertl, Desiree
Hilbring, Lasse Johansson, Kostas Karatzas, Ari Karppinen, Ioannis Kompatsiaris, Tarja
Koskentalo, Simon Mille, Juergen Mossgraber, et al. 2015. “Getting the Environmental

Information across: From the Web to the User.” Expert Systems 32 (3): 405–32.
doi:10.1111/exsy.12100.
Wanner, Leo, Harald Bosch, Nadjet Bouayad-Agha, Gerard Casamayor, Thomas Ertl, Désirée
Hilbring, Lasse Johansson, Kostas Karatzas, Ari Karppinen, Ioannis Kompatsiaris, Tarja
Koskentalo, Simon Mille, Jürgen Moßgraber, et al. 2015. “Getting the Environmental
Information across: From the Web to the User.” Expert Systems 32 (3): 405–32.
doi:10.1111/exsy.12100.
Wilson, Anne, Robert R. Downs, W. Christopher Lenhardt, Carol Meyer, William Michener,
Hampapuram Ramapriyan, and Erin Robinson. 2014. “Realizing the Value of a National Asset:
Scientific Data.” Eos, Transactions American Geophysical Union 95 (50): 477–78.
doi:10.1002/2014EO500006.
Wood, Tim. 2012. “How Are REST APIs Versioned?” Lexical Scope.
http://www.lexicalscope.com/blog/2012/03/12/how-are-rest-apis-versioned/.
Yu, Jonathan, Simon Cox, Gavin Walker, Paul J. Box, and Paul Sheahan. 2011. “Use of Standard
Vocabulary Services in Validation of Water Resources Data Described in XML.” Earth Science
Informatics 4 (3): 125–37.
Yu, Liang, and Yong Liu. 2015. “Using Linked Data in a Heterogeneous Sensor Web: Challenges,
Experiments and Lessons Learned.” International Journal of Digital Earth 8 (1): 15–35.
doi:10.1080/17538947.2013.839007.
Zazueta, Rob. 2015. “The Ultimate Solution to Versioning REST APIs: Content Negotiation.” The
RESTed NARWHL. June. http://www.narwhl.com/2015/03/the-ultimate-solution-to-
versioning-rest-apis-content-negotiation/.
2016. Accessed April 14. http://landing.deloitte.com.au/rs/761-IBL-
328/images/TechTrends2016.pdf.

CONTACT US FOR FURTHER INFORMATION
t 1300 363 400 Peter Taylor
+61 3 9545 2176 Senior Software Engineer
e csiroenquiries@csiro.au t +61 3 62375617
w www.data61.csiro.au e peter.taylor@csiro.au
w www.data61.csiro.au
AT CSIRO WE SHAPE THE FUTURE

We do this by using science and
technology to solve real issues. Our
research makes a difference to industry,
people and the planet.
View publication stats

Environmental WebAPIs State of Art

Uploaded by

Copyright:

Available Formats

Environmental WebAPIs State of Art

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Environmental WebAPIs State of Art

Uploaded by

Copyright:

Available Formats

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

Web APIs for environmental data - state of the art investigation

Technical Report · June 2016

The user has requested enhancement of the downloaded file.

Web APIs for environmental data | 1

Web APIs for environmental data | 2

Web APIs for environmental data | 3

Web APIs for environmental data | 4

Web APIs for environmental data | 5

Web APIs for environmental data | 6

The need for Web API best

Figure 1 - Method for this report

Web APIs for environmental data | 7

Web APIs for environmental data | 8

Primary usage Technologies utilised Core attributes

Enterprise computing XML (XSD, XSLT), W3C  Supports integration of different

Web APIs for environmental data | 9

Research Semantic Web, Linked  Interlinked data across the web

2.1 Why provide APIs?

Web APIs for environmental data | 10

Web APIs for environmental data | 11

2.2 Who are the users of Environmental Web APIs?

Figure 3 - API value chain

Web APIs for environmental data | 12

Blue Sky Industries

Web APIs for environmental data | 13

Web APIs for environmental data | 14

Web APIs for environmental data | 15

Variety Volume Velocity Veracity

Observations, Cheaper more

Points, grids, Push-based

Images, video, Earth Warnings and

Provenance, data Handling 3rd

Figure 5 - Complexities in environmental data

Web APIs for environmental data | 16

Web APIs for environmental data | 17

4.1 National Oceanic and Atmospheric Administration (NOAA)

Web APIs for environmental data | 18

Figure 6 - NOAA NCEI data categories

ONLINE APIS DESCRIPTION FORMAT(S)

Climate API Legacy Station time-series data. XML, CSV

Gridded data service Satellite/Radar/Model Data. NetCDF, XML,

Web APIs for environmental data | 19

Historic Observing Integrated station history database providing in situ JSON

4.1.1 Climate Data API v2.0

•Top level grouping

•A logical grouping of data types

•The instance phenomenon

•Logical grouping of location types

•Individual (point) locations

•Give me the data already

Figure 7 - Climate data online RESTful API overview

Web APIs for environmental data | 20

GET http://www.ncdc.noaa.gov/cdo-web/api/v2/datatypes?limit=1000 (auth header required)

Web APIs for environmental data | 21

GET http://www.ncdc.noaa.gov/cdo-web/api/v2/stations?datatypeid=ANN-PRCP-AVGNDS-GE001HI (auth

GET http://www.ncdc.noaa.gov/cdo-web/api/v2/datasets?datatypeid=ANN-PRCP-AVGNDS-GE001HI (auth

Web APIs for environmental data | 22

Web APIs for environmental data | 23