National Aeronautics and Space Administration

Monitoring and Modeling Floods using Earth Observations

Amita Mehta & Sean McCartney
Guest Speaker: Dr. Augusto Getirana, NASA-GSFC
September 21, 2022
Training Outline
Two 2-hour sessions offered in English with materials available in Spanish

Part 1: September 14, 2022 Part 2: September 21, 2022

Introduction: The Hydrological Modeling
Overview: Flood Monitoring and Modeling
and Analysis Platform (HyMAP)

https://ldas.gsfc.nasa.gov/sites/default/files/ldas/nldas/
https://hydrafloods-servir.adpc.net/map/
presentations/Getirana_NLDAS_HyMAP_10Nov2016.pdf

NASA’s Applied Remote Sensing Training Program 2

Outline for Part 2

• Review of Part 1
• Introduction to The Hydrological Modeling and Analysis Platform
(HyMAP) and its applications
• Summary and Q/A

NASA’s Applied Remote Sensing Training Program 3

Homework and Certificate

• One homework assignment:

– Answers must be submitted via Google Form accessed from the ARSET website
– Homework will be made available on today.
– Due date for homework: October 7, 2022.

• A certificate of completion will be awarded to those who:

– Attend all live webinars
– Complete the homework assignment by the deadline
– You will receive a certificate approximately two months after the completion
of the course from: marines.martins@ssaihq.com

NASA’s Applied Remote Sensing Training Program 4

Review of Part 1
Flood Monitoring Tools

• 1MODIS-Based Flood Mapping (NASA Worldview)

• 1Dartmouth Flood Observatory River Watch (DFO River Watch)
• 2Global Disaster Alert and Coordination System (GDACS)
• Advanced Rapid Imaging and Analysis (ARIA)
• The HYDrologic Remote sensing Analysis for Floods (HYDRAFloods)

1https://appliedsciences.nasa.gov/join-mission/training/satellite-remote-sensing-flood-monitoring-and-management

2https://appliedsciences.nasa.gov/join-mission/training/english/arset-overview-global-disaster-alert-and-coordination-system-gdacs

NASA’s Applied Remote Sensing Training Program 6

Hydrologic and Hydraulic Flood Modeling

• Hydrologic Models: Circulation of

water through the hydrologic
cycle and quantification of runoff
flow produced by precipitation. It
deals with precipitation,
evaporation, infiltration,
groundwater flow, surface runoff,
and streamflow. https://ncar.github.io/hydrology/projects/hydrologic_modeling

• Hydraulics Models: The

mechanical behavior of water in
open or closed channels. It
provides water flow and depth as
water moves from one point to the
next in a channel.
https://dudek.com/do-you-know-the-difference-between-hydrology-and-hydraulics/ https://www.nww.usace.army.mil/Media/Images/igphoto/2002565818/

NASA’s Applied Remote Sensing Training Program 7

Flood Modeling
• Remote sensing observations are routinely • NASA Earth observations used
used for inputs: for flood monitoring, mapping,
- Weather and precipitation data and modeling from:
- Digital elevation
- Land cover
- MODIS
- Landsat
- GPM
- SRTM
- SMAP
- Sentinel-1and -2 (ESA)

NASA’s Applied Remote Sensing Training Program 8

Examples of Open-Source Flood Models
The following models are widely used for flood inundation mapping at the
watershed/river basin scale:
• U.S. Army Corps of Engineers (USACE) • The Soil & Water Assessment Tool (SWAT)
Hydrologic Engineering Center's River
Analysis System (HEC-RAS)

https://ieeexplore.ieee.org/document/8900379 DOI : http://doi.org/10.26480/jcleanwas.02.2018.19.24

NASA’s Applied Remote Sensing Training Program 9

Part 2: NASA Land Information System and HyMAP Routing Model
https://lis.gsfc.nasa.gov/software/lis

https://ldas.gsfc.nasa.gov/sites/default/files/ldas/nldas/presentations/Getirana_NLDAS_HyMAP_10Nov2016.pdf

NASA’s Applied Remote Sensing Training Program 10

Recent progress on flood modeling at NASA Goddard:
From model development to capacity building
Focus: HyMAP Applications
National Aeronautics and Space Administration

Recent progress on flood modeling at NASA Goddard:

From model development to capacity building
Dr. Augusto Getirana

Sep. 21, 2022

 Why should we care about flood modeling?

NASA’s Earth Observatory

NASA’s Applied Remote Sensing Training Program 13

 Climate change and human activities have been magnifying the
occurrence, intensity and impacts of extreme hydrological events

Bangladesh and India (2022) Source: Reuters

Pakistan (2022) Source : Al Jazeera

Kentucky, USA (2022) Source : USA Today

Petropolis, Brazil (2022) Source : NBC

NASA’s Applied Remote Sensing Training Program 14
Satellites help us monitoring extreme events and quantifying their
socioeconomic impacts

August 4, 2022 August 28, 2022

June 1985 July 2022

Lake Powell
Credit: NASA’s Earth Observatory Pakistan

NASA’s Applied Remote Sensing Training Program 16

Source: NASA’s Earth Observatory
 Remote sensing of surface hydrology

Advantages Integration with

o Monitoring of remote regions; computational models
o Spatial distribution; o Evaluation;
o Open access. o Calibration;
o Data assimilation.

Disadvantages
o Temporal resolution; Result
o Limited physical processes; o Estimates of unobserved physical processes;
o Relatively short time series; o Refined spatial and temporal resolutions;
o Past events only. o Scenarios;
o Forecasts.
NASA’s Applied Remote Sensing Training Program 17
 Flood models and river routing schemes
Main characteristics MGB: Vector-based model
•Spatial and temporal scales;
•Discretization;
•Complexity of physical processes.

Global scale
TRIP (Oki et al., 1999)
WaterGAP (Doll et al., 2003)
CTRIP (Decharme et al., 2011)
CaMa-Flood (Yamazaki et al., 2011)
HyMAP (Getirana et al., 2012) HyMAP: Grid-based model
MOSART (Li et al., 2013)
DRTR (Wu et al., 2013)

Continental/meso scale
LISFLOOD-FP (Bates and De Roo, 2000)
MGB (Collischonn et al., 2007)
LEAF-Hydro-Flood (Miguez-Macho et al., 2007)
THMB (Coe et al., 2008)
RAPID (David et al., 2011)

NASA’s Applied Remote Sensing Training Program 18

 LIS surface hydrology capabilities
Implemented and available
Under implementation
Planned for near-future Surface
HyMAP Features Urban
water
TWS DA DA hydro Outputs
Implemented:
• Local inertia formulation
Coastal • Water discharge
• Adaptive time step
Total runoff interface • Water depth
• Backwater effects
LSM `AaAZ • Water storage
• Floodplain dynamics
• 2-way coupling
SWS
• Flow velocity
(2-way coupling) Res. Op.
• Reservoir operation • Floodplain extent
Wat. manag.
• Water management • River slope
Irrigation
• GRACE-DA correction for module
SWS
• Surface water DA (water Lakes and
Bifurcation Forecast
level, discharge and Wat. qual.
extent)
• Urban flood and drainage Flexible spatial scale
• Coastal interface 1o global surface water storage variability 200-m streamflow over
Rio de Janeiro city
Ongoing:
• Bifurcation

Future:
• Sediment dynamics
• Irrigation coupling
• Lake module
 Content of this seminar
• Rivers and floodplains as key components of terrestrial water storage variability

• Climate and human impacts on hydrology and flood risk in Southern Louisiana

• Impacts of dam operation on Lake Victoria, in Eastern Africa

• Development of a flood model for the city of Rio de Janeiro, Brazil

 Content of this seminar
• Rivers and floodplains as key components of terrestrial water storage variability

• Climate and human impacts on hydrology and flood risk in Southern Louisiana

• Impacts of dam operation on Lake Victoria, in Eastern Africa

• Development of a flood model for the city of Rio de Janeiro, Brazil

 Rivers and floodplains as key components of terrestrial
water storage variability

Scientific questions

1. What is the global impact of surface water storage on TWS?

2. How does that impact vary spatially?

NASA’s Applied Remote Sensing Training Program 22

 TWS Modeling Framework
LIS - Surface
LIS
Parameters
LIS - OPT/UE models
Optimization/
Uncertainty Estimation
(GA/MCMC)
Land Surface Models
(Noah, Catchment, VIC)
Lake model (FLake),
Streamflow routing
LIS - DA models (HyMAP) Runoff/Streamflow
Observations Surface fluxes
Data Assimilation
Snowpack
(EnKF) Soil Moisture
yk
x i+
xk− Pk+
1
i+
xik− xk

Pk− LIS - Outputs

t k− 1 tk t k+ 1

Meteorological
Forecasts, Analysis,
Observations

Vertical water and energy

Surface water dynamics; balance;
Source: NASA’s LIS website

Surface water storage Land water storage (LWS)

HyMAP (SWS) Noah-MP

Surface runoff

Baseflow

Getirana et al. (2012, JHM)

NASA’s Applied Remote Sensing Training Program Mendoza et al. (2015, JHM) 23
 Experimental design

Four meteorological forcings and two precipitation datasets, totaling 12 runs;

Meteorological forcings: Precipitation datasets: Modeling protocol:

➢ MERRA2 (1979-present) ➢ TRMM (1998-2016) ➢ 50-year spin up;
➢ GDAS (2000-present) ➢ TRMM-RT (1998-2016) ➢ 1-degree spatial resolution;
➢ Princeton (1948-2014) ➢ 1-hour time step;
➢ ECMWF (2003-present) ➢ 2003-2014 period.

The analysis was performed using the 12-member ensemble mean as the reference.

NASA’s Applied Remote Sensing Training Program 24

GLOBAL TERRESTRIAL WATER STORAGE
➢ About 2,860km3 of water is stored in rivers and floodplains
[previous estimates vary from 2,000km3 (Oki & Kanae, 2006) to
2,120km3 (Shiklomanov, 1993)];
➢ 28% of that water, or 800km3, is concentrated in the Amazon
Basin;
➢ Uncertainty due to meteorological forcings is about 6% globally.
 1
IMPACTS ON TWS
0.5
SWS=8% globally
0

SWS impact

1 1

0.5 0.5

0 0

GWS impact SWS impact

1 1 1

0.5 0.5 0.5

0 0 0

SWS impact SM impact GWS impact

1 1 1

0.5 0.5 0.5

0 0 0

GWS impact SWE impact SM impact

GWS=17% globally SWE=23% globally SM=51% globally
• Rivers and floodplains as key components of terrestrial water storage variability

• Climate and human impacts on hydrology and flood risk in Southern Louisiana

• Impacts of dam operation on Lake Victoria, in Eastern Africa

• Development of a flood model for the city of Rio de Janeiro, Brazil

 Climate and human impacts on hydrology and flood risk in
Southern Louisiana
Observation-based evidence of hydrological change, sea level rise, increased terrestrial water storage and land loss

NASA’s Applied Remote Sensing Training Program 28

 MRD model configuration
• 1993-2020 period;
River geometry parameterization
• 2km spatial resolution;
• MERRA-2;
• Streamflow obs used as upstream boundary
condition at 3 locations;
• Radar altimetry used as downstream
boundary condition;
• 4 experiments;
Scenario Hydrology Sea level Water
management
1 Yes 7-day Yes
2 Yes 7-day No
3 Yes Climatology No
4 Yes Constant No
• Representation of water management

Old
River Morganza
Floodway
Lake
Pontchartrain

Bonnet Carre
Spillway
 Model-based MRD hydrological changes
Based on Scenario 1 (i.e., all factors considered), surface elevation and flooded area generally
increase over the domain from 1993-97 to 2016-20

NASA’s Applied Remote Sensing Training Program 30

 Coastal interface & water management
3 factors considered in our modeling system: Climate-induced hydrological changes (CHC), sea level rise (SLR) and water
management

Cropland flood risk [km2 ] Cropland flood risk [km ] Population flood risk [residents]
80 80 40,000

40 40 20,000

0 0 0
SLR CHC Water All
SLRfactors CHC SLR
Water CHC Water All factors
management management management
-40 -40 -20,000

(a) (b)
-80 -80 -40,000
Median event Annual event
Median event Annual event Median event Annual event

Flood risk over coastal Louisiana as a function of return period (median and annual
events) due to SLR, CHC, water management and all factors combined (all factors).

Changes in flooded areas due to CHC, SLR, and water

management NASA’s Applied Remote Sensing Training Program
over coastal Louisiana for annual events 31
• Rivers and floodplains as key components of terrestrial water storage variability

• Climate and human impacts on hydrology and flood risk in Southern Louisiana

• Impacts of dam operation on Lake Victoria, in Eastern Africa

• Development of a flood model for the city of Rio de Janeiro, Brazil

Impacts of dam operation on Lake Victoria, in Eastern Africa

Objectives
1. Propose and implement a satellite-based reservoir operation
scheme in HyMAP;

2. Evaluate impacts of such a scheme on the representation of:

i. surface water dynamics (water elevation and extent);
ii. terrestrial water storage variability.
NASA’s Applied Remote Sensing Training Program 33
 Implementation of a satellite-based reservoir operation scheme

Satellite-based reservoir operation scheme

Qi Qi Qi Qi
Hi Hi Hi Hi
dS S’ H’o H’o
Hobs S’
S Qo S Qo
Q’o Q’o
Ho Ho

If (Qo – Qmin) x dt ≥ dS If (Qo – Qmin) x dt < dS

Yangtze
Three River
Gorges Three Gorges
1992)
(Google Earth 2019) (Google Earth 2019)

Grid cell Grid cell

(naturalized system) (Reservoir operation)
 Lake Victoria • Located in Eastern Africa;
• Shared by Uganda, Tanzania and Kenya;
• World’s second largest freshwater lake after
Lake Superior in the US.;
• Water extent: ~66,500km2;
• Water volume: ~2,750km2;
• Drainage area: ~252,000km2;
• White Nile River headwaters;
• Directly supports the livelihood of more than
42 million people;
• Controlled since 1954 (Nalubaale dam).

Parameterization
Trans-sectional profile
• 3-arcsec MERIT Hydro (Yamazaki et al., 2017): Dam
DEM, flow direction, drainage area;
• 3-arcsec bathymetry (Hamilton et al., 2016);
• Hydroweb’s radar altimetry time series (Cretaux
et al., 2011).
Jason track
 Experimental design
❑ Hydrological models:
• Noah-MP land surface model;
• Representation of groundwater;

Noah-MP HyMAP • HyMAP global flood model;

• Local inertia formulation;
• Adaptive time step.
Surface runoff

Baseflow ❑ Experiments:
• Naturalized system;
• Reservoir operation;
• Qmin=0;
• 0.1-degree spatial resolution;
• 15-min time step;
• 2003-present.
Mendoza et al. (2015, JHM) Getirana et al. (2012, JHM)

❑ Meteorological forcings:
• MERRA-2 (M2; 1979-present) ;
• CHIRPS (M2C; 1981-present);
• ECMWF (2003-present);
• GDAS (2000-present).
NASA’s Applied Remote Sensing Training Program 36
 NS: -326.6 0.91
1130
r: 0.38 0.97 RESULTS
: 0.38 1.00
Lake water elevation [m] M2 M2C ECMWF GDAS
1129
1135
RMSE 0.08 0.10 0.17 0.05
1128
1134
NS 0.95 0.93 0.79 0.99
1127
1133 r 0.98 0.97 0.94 0.99

1126 𝛾 0.99 0.97 0.93 1.00

1132
Issues filling the lake
1125
1131 • All forcings present good metrics for water
RMSE:
2004 6.62 2008
0.09 7-m
2012difference 2016 elevation;
NS: -311.6 0.94 • GDAS shows the best overall result;
1130
r: 0.53
Natural 0.98
Reservoir Radar Altimetry • There are issues filling the lake, in particular,
1129 : 0.52 1.03 GDAS ECMWF;
• There is about 7-m difference between
1128 experiments with reservoir and naturalized
system;
1127
• GDAS has too much water and ECMWF has
1126 too little;
ECMWF • Ensemble median is used for comparison;
1125 • All metrics are significantly improved with the
2004 2008 2012 2016 reservoir operation scheme.

Natural Reservoir
NASA’s Applied Remote Sensing Training Program Radar Altimetry 37
 WATER EXTENT
Lake Kyoga

Lake water extent [10 33km 22]

68

67

66

65
Mean: 60.9 66.4 66.8 66.6 • Long-term average modeled water extents match satellite
: 0.58 0.47 0.14 0.46 observations; (𝜇 = mean)
64
[km2/year]: 64 64 -13 34
• Reservoir operation and Landsat-based variability (standard
63 deviation) are similar;
• Except for MODIS, all trends are positive (both model experiments
62
show steeper trends);
61 • Water extent derived from reservoir operation shows the best
agreement with radar altimetry.
60

2004 2008 2012 2016

2016

Natural Natural
Reservoir Reservoir
MODIS Landsat

NASA’s Applied Remote Sensing Training Program 38

• Rivers and floodplains as key components of terrestrial water storage variability

• Climate and human impacts on hydrology and flood risk in Southern Louisiana

• Impacts of dam operation on Lake Victoria, in Eastern Africa

• Development of a flood model for the city of Rio de Janeiro, Brazil

NASA-Rio de Janeiro Partnership

Courtesy: D. Kirschbaum
Goal and proposed plan NASA tools, Local expertise and
models and datasets monitoring capabilities

• Land Information System (LIS);

• Land surface Data Toolkit (LDT);
• GPM;
• Landsat; Operational
• Modeling and satellite expertise. customized
• Real-time hydrometeorological monitoring; flood
• Local land use and soil type distribution; monitoring
• Lidar-based DEM; system
• River network and geometry;
• Flood mapping;
• Urban drainage system, etc.

• Frequent phone calls, emails; Capacity building

Engagement, knowledge exchange, training,
• Updates on results, challenges and requirements; outreach
• Annual in person meetings and workshops;
• Strong local interest and technical expertise;
• Easy and immediate access to information, etc.
HyMAP’s urban drainage module

Drainage network parameters

Street block length [m] 100
Underground squared pipe width [m] 1
Urban drainage map Drainage length density [m/m2] 0.02
Roughness coefficient for cement pipes [-] 0.012
WG Drainage system slope [m/m] 0.003
HG

Gutter parameters
Gutter width [m] 1
Sidewalk Sidewalk

Gutter Street Gutter

Gutter height [m] 0.15

Gutter density [units/m2] 0.0008

Gutter water intake velocity [m/s] 2

Urban drainage
system
 Preliminary results
Ilha do Fundão

Manguinhos
São Cristovão

Downtown

Downscaled water depth

anomalies at 10-meter spatial
resolution during major flood
occurred in April 2010

Lagoa

Water depth anomalies using Leblon/Ipanema

00:00Z 04-04-2010 as baseline
• Rivers and floodplains as key components of terrestrial water storage variability

• Climate and human impacts on hydrology and flood risk in Southern Louisiana

• Impacts of dam operation on Lake Victoria, in Eastern Africa

• Development of a flood model for the city of Rio de Janeiro, Brazil

 Final considerations & prospects
• HyMAP has been significantly improved in the past 11 years, since its initial implementation in
LIS;

• The combination of all capabilities currently available in HyMAP puts it as one of the most
advanced global flood models available;

• Ongoing and future developments:

- Bifurcation;
- Improved reservoir operation scheme;
- Inclusion of natural lakes and a water quality module.

• HyMAP is freely available in LIS through GitHub (https://github.com/NASA-LIS/LISF);

• HyMAP outputs are available from FLDAS, NCA-LDAS, and soon from GLDAS
(https://ldas.gsfc.nasa.gov/);

• New collaborations for either new developments and applications are always welcome!

NASA’s Applied Remote Sensing Training Program 45

Summary
Flood Monitoring and Mapping
Passive Microwave Observations-based
MODIS Flood Map River Discharge

https://worldview.earthdata.nasa.gov/ https://floodobservatory.colorado.edu/DischargeA
ccess.html
Sentinel-1 SAR-based flood Mapping
Multiple Data Sources
ARIA
GDACS

https://aria.jpl.nasa.gov/index.html
https://gdacs.org/
NASA’s Applied Remote Sensing Training Program 47
HYDrologic Remote Sensing Analysis for Floods (HYDRAFloods)
https://servir-mekong.github.io/hydra-floods/algorithms/

• HYDRAFloods uses the following • HYDRAFloods allows access to

remote sensing inputs: high-quality, cloud-based surface
- Sentinel-1 C-band SAR data water mapping algorithms:
- Sentinel-2 optical data - SAR Speckle Filtering Algorithms

- Landsat 7 & 8 optical data - Correction Algorithms

- VIIRS and MODIS optical data - Applying Illumination Correction on

Optical Imagery
• A publicly available, web-based, near - Applying Slope Correction on SAR
real-time flood monitoring tool Imagery
• Developed using Google Earth Engine - Generic Water Mapping
and Google Cloud Platform with Python Algorithms
API.

NASA’s Applied Remote Sensing Training Program 48

Examples of Flood Models

HEC-RAS SWAT

https://www.hec.usace.army.mil/software/hec-ras/ https://swat.tamu.edu/

NASA’s Applied Remote Sensing Training Program 49

Flood Modeling System (LIS-HyMAP)

• HyMAP is freely available in LIS

through GitHub
(https://github.com/NASA-
LIS/LISF);

• HyMAP outputs are available

from FLDAS, NCA-LDAS, and
soon from GLDAS
(https://ldas.gsfc.nasa.gov/);

From: A. Getrirana

NASA’s Applied Remote Sensing Training Program 50

Questions?

• Please enter your questions in Bangladesh Floods – June 2022

the Q&A box. We will answer
them in the order they were
received.

• We will post the Q&A to the

training website following the
conclusion of the webinar.

https://maps.disasters.nasa.gov/arcgis/home/item.html
?id=9e49f039612741878e7f560a67f2d4e5

NASA’s Applied Remote Sensing Training Program 51

Contacts
• Trainers:
– Augusto Getirana: augusto.getirana@nasa.gov
– Amita Mehta: amita.v.mehta@nasa.gov
– Sean McCartney: sean.mccartney@nasa.gov
• Training Webpage:
– https://appliedsciences.nasa.gov/join-
mission/training/english/arset-monitoring-and-modeling-
floods-using-earth-observations
• ARSET Website: Check out our sister programs:

– https://appliedsciences.nasa.gov/arset
• Twitter: @NASAARSET

NASA’s Applied Remote Sensing Training Program 52

 Thank You!

NASA’s Applied Remote Sensing Training Program 53