Habitat Modelling

Correlative species distribution models are becoming commonplace in the scientific literature and public outreach
products, displaying locations, abundance, or suitable environmental conditions for harmful invasive
species, threatened and endangered species, or species of special concern. Accurate species distribution models
are useful for efficient and adaptive management and conservation, research, and ecological forecasting. Yet,
these models are often presented without fully examining or explaining the caveats for their proper use and
interpretation and are often implemented without understanding the limitations and assumptions of the
model being used. We describe common pitfalls, assumptions, and caveats of correlative species distribution
models to help novice users and end users better interpret these models. Four primary caveats corresponding
to different phases of the modeling process, each with supporting documentation and examples, include:
(1) all sampling data are incomplete and potentially biased; (2) predictor variables must capture distribution
constraints; (3) no single model works best for all species, in all areas, at all spatial scales, and over time; and
(4) the results of species distribution models should be treated like a hypothesis to be tested and validated
with additional sampling and modeling in an iterative process.

Systematic conservation planning (SCP) to design marine protected areas (MPAs) has traditionally focused on species distributions or benthic habitat features that drive the determination of conservation priorities. Pelagic ecosystem protection is usually incidental because these ecosystems are often data-poor and are difficult to visualize in a planning context. Pelagic ecosystems, however, face increasing and cumulative impacts from threats such as overfishing and climate change, and a precautionary approach is required to protect both known and unknown biodiversity patterns and ecosystem processes. Data-driven pelagic habitat classifications are important when planning for habitat protection in the absence of sufficient in-situ data. In this study, we describe a method for creating a bioregional map of the upper-mixed layer of South Africa's pelagic realm. We selected relevant variables and parameters that best reflect key ecosystem properties at broad, meso, and local scales. We conducted a hierarchal cluster analysis using open-access sea surface temperature (SST), chlorophyll-a (chl-a), net primary productivity (NPP), mean sea level anomalies (MSLA), and seabed slope and depth data. The resulting map delineates three bioregions subdivided into seven biozones and sixteen pelagic habitats within South Africa's continental Exclusive Economic Zone (EEZ). This habitat map was incorporated into SCP of a proposed expanded MPA network that includes offshore protected areas and meets National objectives. The proposed network will increase protection of the pelagic realm (>30 m depth) of the EEZ from 0.002% to 6.0%. We contend that bioregional analyses based on publicly available remote-sensing data are useful for identification of offshore habitats, especially when robust biological data are unavailable, as a framework for ecosystem reporting, and for inclusion in a systematic design for a representative offshore MPA network. Further research should focus on modelling and mapping the permanence of pelagic habitats and different spatio-temporal scales of variability, validating habitat boundaries with biological data, and understanding the threats and efficacy of achieving pelagic protection through management mechanisms like MPAs.

Networks of Marine Protected Areas are implemented to conserve fish populations, yet their effectiveness is rarely comprehensively examined or adaptively managed. In this dissertation, I evaluate a network of Rockfish Conservation Areas (RCAs) implemented to reverse population declines of inshore Pacific rockfishes (Sebastes spp.).
First, I used SCUBA surveys to examine patterns of Black Rockfish abundance compared to spatial and temporal variability in recruitment to determine how recruitment influences population density in and around a RCA. Habitat variables such as complexity and rocky substrate predicted adult Black Rockfish abundance while recruitment did not.
Next, I surveyed the fish communities of 35 RCAs and adjacent unprotected areas using a Remotely Operated Vehicle (ROV). Habitat features such as percent rocky substrates and depth influenced the density of Quillback (S. maliger), Yelloweye (S. ruberrimus), Greenstriped Rockfishes (S. elongatus), Kelp Greenling (Hexagrammos decagrammus), Lingcod (Ophiodon elongatus) and all inshore rockfishes combined, while reserve status did not. The results give little indication that demersal fish populations have recovered inside the RCA system.
I used aerial observations of recreational fishing from surveys before, during and after 77 RCAs were established and found there was no evidence of a change in fishing effort in 83% of the RCAs. Compliance was related to the level of fishing effort around the RCA, the size and perimeter-to-area ratio of RCAs, proximity to fishing lodges and the level of enforcement. Non-compliance in RCAs may be hampering their effectiveness and impeding rockfish recovery.
Lastly, I modeled rocky reef habitat using Random Forest Classification to assess habitat in the RCAs. I combined three habitat metrics with data on compliance, RCA size, rockfish bycatch, and connectivity into a single Conservation Score. The Conservation Score is related to the log reserve ratio, a measure of relative abundance, for Quillback Rockfish. RCAs with low Conservation Scores are not likely to be effective and managers should evaluate the reasons for low scores and address reserve shortcomings in an adaptive spatial management framework. Education and enforcement efforts are critical to the recovery of depleted fish stocks. Continued monitoring and evaluation of the RCAs is essential to rockfish conservation.

Man-made influences on morphological and hydrological processes in the last decades have lead to substantial changes in river characteristics. The hydrologly, morphological heterogeneity, water quality, substrate characteristics and in some cases the entire river continuum were completely modified leading to the impacts on river ecology which we have today. Thus, in river management models are reqiored which are able to quantify the species-environment relationships within aquatic habitats. Aquatic habitat simulation models (AHSMs) allow for such quantification because they predict species distributions based on data which describes the abiotic environment. This article presents a comparison between a univariate approach based on preference functions and a knowledge-based multivariate fuzzy- logical approach in the River Aare, Switzerland. In particular, the different methods used to link abiotic environmental parameters to biological responses are considered. Given the results, the approach based on preference functions lacks a proper weighting and combination of the separate habitat suitabilities. As two preference functions were necessary to describe the study site, the transferability of habitat requirements for preference functions is very limited due to the crisp boundaries for the definition of habitat requirements. As ecological transitions are more gradual, the fuzzy-logical approach may serve to be a more appropriate modeling technique to deal with these ecological gradients by using imprecise or qualitative information numerically in form of overlapping fuzzy-sets. Given to these advantages the multivariate fuzzy-logical approach is found to be the more robust approach when compared to univariate preference functions.

"Machine learning (ML) techniques have become important to support decision making in management and conservation of freshwater aquatic ecosystems. Given the large number of ML techniques and to improve the understanding of ML utility in ecology, it is necessary to perform comparative studies of these techniques as a preparatory analysis for future model applications. The objectives of this study were (i) to compare the reliability and ecological relevance of two predictive models for fish richness, based on the techniques of artificial neural networks (ANN) and random forests (RF) and (ii) to evaluate the conformity in terms of selected important variables between the two modelling approaches. The effectiveness of the models were evaluated using three performance metrics: the determination coefficient (R2), the mean squared error (MSE) and the adjusted determination coefficient (R2adj) and both models were developed using a k-fold crossvalidation procedure. According to the results, both techniques had similar validation performance (R2 = 68% for RF and R2 = 66% for ANN).
Although the two methods selected different subsets of input variables, both models demonstrated high ecological relevance for the conservation of native fish in the Mediterranean region. Moreover, this work shows how the use of different modelling methods can assist the critical analysis of predictions at a catchment scale. Copyright ONEMA, 2013. "

This extension note is the eighth in a series of eight that describes a set of tools and processes developed to support sustainable forest management planning and its pilot application in the Arrow Timber Supply Area (TSA). It summarizes the criterion and indicators used to evaluate quality-of-life opportunities for the sustainable forest management (SFM) pilot basecase analysis of the Lemon Landscape Unit. The management of forests has broadened to include various social values and amenities that were considered during the development of criteria and indicators for the Arrow Innovative Forestry Practices Agreement. The quality-of-life criterion was assessed through indicators that addressed outdoor recreation opportuni-ties and visual quality of the managed landscape. This assessment was informed by public input from area residents and stakeholders. Measurable quality-of-life indicators allowed trade-offs with other resources in the SFM pilot basecase analysis; protection of these ...

Asian elephants are isolated in fragmented habitat patches in and around Bardia National Park (BNP), Nepal. To describe habitat use patterns and ecogeographical variables (EGVs) that determine an elephant's niche in BNP, we used a General Niche-Environment System Factor Analysis (GNESFA) modeling framework. Novel to our study was the comparison of niche requirements between core (residential) and corridor (travel corridor) areas to elucidate site-specific preferences of Asian elephants in BNP. A total of 13 EGVs (four topographic variables, six land covers, heterogeneity index and two anthropogenic variables)
were examined. We implemented a ‘bias file’ approach to address potential sampling bias in the transect survey methods for presence records. Our study illustrated that, regardless of study area, elephants' habitat use was positively influenced by presence of grasslands, mixed forest, and landscape heterogeneity, whereas use was restricted by the
topographic variables of slope and elevation. Results also demonstrated different habitat preferences between elephants in the core and corridor, which may be attributed to differences in potential dangers posed in these areas; in the core, elephant habitat preference was mainly associated with food resources such as grassland or mixed forest, whereas in the corridor, where elephants are more likely to encounter human conflict, the anthropogenic factor of distance to human settlements contributed the most in predicting elephant presence. Correlations among significant factors from the three methods (FANTER,
ENFA, and MADIFA) demonstrated the consistent and reliable results of these approaches. While these methods complemented each other by providing different points of view, FANTER was especially useful when bimodal niches were analyzed. We suggest a detailed conservation plan for the small populations of elephants in BNP and surrounding areas, while considering the protection of travel routes from human activities in the corridor habitats, and lastly, maintaining grasslands and waterholes in core habitats.

Knowledge of the relationships between species and
environmental variables requires understanding the factors influencing
the distribution patterns of populations. The Southern Brazilian coast
is an important reproductive and calving site for Southern right whale
Eubalaena australis (Desmoulins, 1822) (Cetartiodactyla, Balaenidae),
which currently shows population growth, and variability in the
number of individuals migrating, annually, to this area. This study
investigated the hypothesis that the environmental and oceanographic
factors ENSO (El Niño Southern Oscillation), SST (Sea Surface
Temperature), shelf slope and amount of bays, influence southern right
whales concentrating in specific areas, as well as the interannual
variability in the number of individuals. Right whales data consisted of
number and localization of individuals, sighted through helicopter
expeditions between 27.4° and 30.4°S. Firstly, the population spatial
distribution was analyzed through density estimative maps.
Relationships between interannual variability in the number of right
whales and ENSO were tested using nonparametric methods.
Generalized Additive Models (GAMs) were applied for testing
nonlinear relationships between right whales and the covariates SST,
shoreline embayment index, shelf slope and latitude. Spatial
distribution analysis showed higher concentration of right whales
between Imbituba and Santa Marta Cape, which is a region
characterized by cooler waters, many bays and higher shelf slope.
Among the correlations tested between ENSO and right whales, it was
significant at 0.05 only for the Nino4 index, on the feeding season,
within a six years lag. GAMs results showed higher number of right
whales around 28.5°S, SST between 15 and 18°C, higher number of
bays and higher shelf slope, highlighting the importance of these
environmental factors for habitat selection in right whales on the
Southern Brazilian coast.

I examined biodiversity management issues, implications and opportunities in a working industrial landscape of the San Bernardino Mountains, California, USA, where historic and active mining operations overlap habitat occupied by threatened, endangered or protected plant and animal tax. To elucidate patterns of habitat use by ungulates in a mini-influenced landscape, I evaluated resource selection by the Cushenbury population of desert bighorn sheep (Ovis Canadensis nelsoni). Modeling of a resource selection function (RSF) for that population identified proximity to active mine areas, water sources, and revegetation sites as important determinants of habitat selection. Next, I reviewed two collaborative dialogues between the public, private and civil sectors resulting in workable solutions achieving biodiversity conservation goals simultaneous with economic development: a habitat management strategy for threatened and endangered plants and a research collaborative focused on the long-term viability of an isolated bighorn sheep population. Finally, I integrated results of RSF modeling with principles adopted by those two initiatives into a conceptual framework to aid the development of an adaptive management plan for the Cushenbury bighorn sheep population. A formulation of conservation value is presented using results of RSF analysis and mitigation credits reflecting the degree to which degraded habitat is enhanced to benefit wild sheep. The proposed Bighorn Habitat Assessment Tool (BHAT) seeks to a) establish a habitat reserve providing maximum benefit to the unique requirements of bighorn sheep, b) incentivize voluntary actions by industry to ensure mining activities are compatible with bighorn sheep conservation, and c) allow for the objective evaluation of multiple mine planning and resource management alternatives.

The determination of Ce, Co, Cr, Cs, Eu, Fe, Gd, Hf, Mn, Rb, Sb, Sc, Ta, Tb, Th, Zn, Zr and Yb by instrumental neutron activation analysis in sediment samples collected from the Gulf of Gaeta (Central Tyrrhenian Sea) is described. The granulometric fractions of the sediments were grouped by cluster analysis into three granulometric facies (sand, clayey silt, and

Asphodelus bento-rainhae P. Silva is a plant species belonging to the order of Liliales. This endemic plant occurs in central Portugal in an area of only 700 hectare on the northern slopes of the Serra da Gardunha (Fundão, Portugal). It is mainly present in north to northeast facing slopes and found in deep soils and open areas, where it is found at higher densities. It occurs along the edges of oak (Quercus robur and Quercus pyrenaica) and chestnut (Castanea sativa) forests, often reaching the herbaceous edges of these woods. Sometimes, it survives in pine forests of Pinus pinaster, mixed woods and slopes or roadsides and in lower abundances in scrubby vegetation. This species survival is threatened by forest fires and the conversion of woodland into cherry orchards, and the fruit growers’ extensive use of herbicides. The tree plantations, as well as urban sprawl, have decreased the extent and quality of its habitat. It is also confronted with invasive plants, such as Acacia dealbata. The mapping of the A. bento-rainhae habitat suitability index results on the integration of a set of biophysical factors using the ArcGIS 10.3 software, based on the Sectorial Plan for Natura 2000 Network and other literature concerning the species ecology. The data used included bioclimatic, soil and topographic variables, based on a digital terrain model (DTM). All the geographic themes (criteria) were classified into three suitability levels: unfavorable, less favorable and favorable. Based on the resulting reclassified themes, the Habitat Suitability Index (HSI) for A. bento-rainhae P. Silva was calculated using the Analytic Hierarchy Process (AHP). The fundamental concept of AHP lies in proceeding from a pairwise comparison of criteria to evaluate the weights that assign relative importance to these criteria. In the end, a map algebra was performed in order to obtain the final theme representing the habitat suitability for A. bento-rainhae. The results regarding the actual species distribution, obtained in the aim of fied work performed in the LIFE-Nature project “Asphodelus bento-rainhae - Measures to manage and preserve it" show a high correlation with the suitability values.