Predicting Broad-scale Patterns of Avian Biodiversity with Landsat Image Texture

Birds are closely associated with habitat structure. Patrick Culbert is using satellite image texture to characterize vegetation structure in the contiguous United States and predict bird species richness wall-to-wall, thus filling the gaps between the places where the North American Breeding Bird Survey is conducted annually.

There is a global concern about loss of species and biodiversity. One of the main causes of species declines is the loss of habitat. In order to make efficient conservation plans it is important to have accurate maps of species richness. For birds, vegetation structure is known to affect species richness. One method for measure bird species richness in small areas is measuring habitat structure in the field. However, such field measurements are not feasible across large areas, such as the contiguous United States, and that is why good biodiversity maps are lacking.

Patrick Culbert is using measures of texture from satellite images to quantify habitat structure without taking field measurements. Image texture is a method to measure the roughness and unevenness of a satellite image in a particular area. For example, agricultural crops will look uniform in image texture as the vegetation composition is homogeneous, while a forest with different stages of succession will look heterogeneous in image texture.

Landsat_Texture_RSE_small_for_Web.jpg

(Left) Landsat (band 4) imagery for Midwestern United States. (Right) Homogeneity for band 4.

In order to relate the texture measurements with bird species richness, Patrick is using the North American Breeding Bird Survey (BBS). BBS consist of about 3000 routes in the United States. Each 24.5 mile route is surveyed annually, with a three-minute count every half mile, counting all birds seen or heard. Around each BBS route, eight image texture measurements; like contrast, and homogeneity, are analyzed, using statistical models to relate image texture with bird species richness. Those results will help to predict bird species richness in areas where the surveys have not been conducted, by analyzing its image texture.

Story By: Anonymous