Linking animal movement and remote sensing – mapping resource suitability from a remote sensing perspective

Remelgado, Ruben; Leutner, Benjamin; Safi, Kamran; Sonnenschein, Ruth; Kuebert, Carina; Wegmann, Martin

doi:10.1002/rse2.70

Cited by 24 publications

(36 citation statements)

References 78 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Remelgado et al. () previously used seasonal Landsat surface reflectance composites for modeling resource availability for white storks ( Ciconia ciconia ). Our study showcases how spectral‐temporal metrics from multiple spectral indices can be linked to different habitat aspects, allowing for a comprehensive and meaningful characterization of wildlife habitat from medium resolution imagery.…”

Section: Discussionmentioning

confidence: 99%

“…Adapting the temporal windows to local phenology (Remelgado et al. ) or using metric definitions that allow comparing values across different phenological curves (e.g. highest/lowest quarterly median; Coops et al.…”

Section: Discussionmentioning

confidence: 99%

“…; Remelgado et al. ). For defining the temporal windows, we used the day of year (DOY) of image recording to categorize Landsat images into start‐of‐season (DOY 60‐151, 42 images), peak‐of‐season (DOY 152‐243, 53 images), and end‐of‐season observations (DOY 244‐334, 54 images).…”

Section: Methodsmentioning

confidence: 98%

“…; Remelgado et al. ). Moreover, the Landsat archive provides freely available, global imagery back to the 1970s (Wulder et al.…”

Section: Introductionmentioning

confidence: 98%

See 3 more Smart Citations

Habitat metrics based on multi‐temporal Landsat imagery for mapping large mammal habitat

Oeser

Heurich

Senf

et al. 2019

Remote Sens Ecol Conserv

View full text Add to dashboard Cite

Up‐to‐date and fine‐scale habitat information is essential for managing and conserving wildlife. Studies assessing wildlife habitat commonly rely on categorical land‐cover maps as predictors in habitat models. However, broad land‐cover categories often do not adequately capture key habitat features and generating robust land‐cover maps is challenging and laborious. Continuous variables derived directly from satellite imagery provide an alternative for capturing land‐cover characteristics in habitat models. Improved data availability and processing capacities now allow integrating all available images from medium‐resolution sensors in compositing approaches that derive spectral‐temporal metrics at the pixel level, summarizing spectral responses over time. In this study, we assessed the usefulness of such metrics derived from Landsat imagery for mapping wildlife habitat. We categorize spectral‐temporal metrics into habitat metrics characterizing different aspects of wildlife habitat. Comparing the performance of these metrics against categorical land‐cover maps in habitat models for lynx, red deer and roe deer, we found that models using habitat metrics consistently outperformed models based on categorical land‐cover maps, with average improvements of 13.7% in model AUC and 9.7% in the Continuous Boyce Index. Performance increases were larger for seasonal habitat models, indicating that the habitat metrics capture intra‐annual variability in habitat conditions better than land‐cover maps. Comparing suitability maps to ancillary data further revealed that our habitat metrics were sensitive to fine‐scale heterogeneity in habitat associated with forest structure. Overall, our study highlights the considerable potential of Landsat‐based spectral temporal metrics for assessing wildlife habitat. Given these metrics can be derived directly and in an automatized fashion from globally and freely available Landsat imagery, they open up new possibilities for monitoring habitat dynamics in space and time.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Discussionmentioning

confidence: 99%

Section: Methodsmentioning

confidence: 98%

“…; Remelgado et al. ). Moreover, the Landsat archive provides freely available, global imagery back to the 1970s (Wulder et al.…”

Section: Introductionmentioning

confidence: 98%

See 2 more Smart Citations

Habitat metrics based on multi‐temporal Landsat imagery for mapping large mammal habitat

Oeser

Heurich

Senf

et al. 2019

Remote Sens Ecol Conserv

View full text Add to dashboard Cite

show abstract

“…To select absences, we used the algorithm proposed in a recent paper (Remelgado et al. ). This algorithm identifies spatially connected groups of samples associated to presences and performs a Principle Components Analysis (PCA) over the environmental conditions associated to the sampled pixels – here described by the NDVI time series.…”

Section: Methodsmentioning

confidence: 99%

From ecology to remote sensing: using animals to map land cover

Remelgado

Safi

Wegmann³

2019

Remote Sens Ecol Conserv

Self Cite

View full text Add to dashboard Cite

Land cover is a key variable in monitoring applications and new processing technologies made deriving this information easier. Yet, classification algorithms remain dependent on samples collected on the field and field campaigns are limited by financial, infrastructural and political boundaries. Here, animal tracking data could be an asset. Looking at the land cover dependencies of animal behaviour, we can obtain land cover samples over places that are difficult to access. Following this premise, we evaluated the potential of animal movement data to map land cover. Specifically, we used 13 White Storks (Cicona cicona) individuals of the same population to map agriculture within three test regions distributed along their migratory track. The White Stork has adapted to foraging over agricultural lands, making it an ideal source of samples to map this land use. We applied a presence–absence modelling approach over a Normalized Difference Vegetation Index (NDVI) time series and validated our classifications, with high‐resolution land cover information. Our results suggest White Stork movement is useful to map agriculture, however, we identified some limitations. We achieved high accuracies (F1‐scores > 0.8) for two test regions, but observed poor results over one region. This can be explained by differences in land management practices. The animals preferred agriculture in every test region, but our data showed a biased distribution of training samples between irrigated and non‐irrigated land. When both options occurred, the animals disregarded non‐irrigated land leading to its misclassification as non‐agriculture. Additionally, we found difference between the GPS observation dates and the harvest times for non‐irrigated crops. Given the White Stork takes advantage of managed land to search for prey, the inactivity of these fields was the likely culprit of their underrepresentation. Including more species attracted to agriculture – with other land‐use dependencies and observation times – can contribute to better results in similar applications.

show abstract

Day versus night use of forest by red and roe deer as determined by Corine Land Cover and Copernicus Tree Cover Density: assessing use of geographic layers in movement ecology

et al. 2022

View full text Add to dashboard Cite

Context Diel use of forest and open habitats by large herbivores is linked to species-specific needs of multiple and heterogeneous resources. However, forest cover layers might deviate considerably for a given landscape, potentially affecting evaluations of animals’ habitat use. Objectives We assessed inconsistency in the estimates of diel forest use by red and roe deer at GPS location and home range (HR) levels, using two geographic layers: Tree Cover Density (TCD) and Corine Land Cover (CLC). Methods We first measured the classification mismatch of red and roe deer GPS locations between TCD and CLC, also with respect to habitat units’ size. Then, we used Generalised Least Squares models to assess the proportional use of forest at day and night at the GPS location and HR levels, both with TCD and CLC. Results About 20% of the GPS locations were inconsistently classified as forest or open habitat by the two layers, particularly within smaller habitat units. Overall proportion of forest and open habitat, though, was very similar for both layers. In all populations, both deer species used forest more at day than at night and this pattern was more evident with TCD than with CLC. However, at the HR level, forest use estimates were only marginally different between the two layers. Conclusions When estimating animal habitat use, geographic layer choice requires careful evaluation with respect to ecological questions and target species. Habitat use analyses based on GPS locations are more sensitive to layer choice than those based on home ranges.

show abstract

Linking animal movement and remote sensing – mapping resource suitability from a remote sensing perspective

Cited by 24 publications

References 78 publications

Habitat metrics based on multi‐temporal Landsat imagery for mapping large mammal habitat

Habitat metrics based on multi‐temporal Landsat imagery for mapping large mammal habitat

From ecology to remote sensing: using animals to map land cover

Day versus night use of forest by red and roe deer as determined by Corine Land Cover and Copernicus Tree Cover Density: assessing use of geographic layers in movement ecology

Contact Info

Product

Resources

About