The response of eastern Ontario farmland species diversity to landscape dynamics was modeled at multiple spatial and temporal scales using metrics derived from MODIS and Landsat data in random forest and multiple regression analyses. For the seven taxa studied, significant positive gamma diversity responses to MODIS-derived start of season NDVI and its spatial coefficient of variation over 3 km x 3 km landscapes were found for beetles and birds. Butterfly diversity was positively related to mean growing season NDVI. Plant diversity was found to be greater in landscapes where the start of the growing season is earlier. Landsat data analysis found that beetle and plant diversity is lower when Tasseled Cap brightness increases and wetness decreases, while butterfly and bird diversity increased with increasing NDVI. Initial inter-annual analysis showed that variation in NDVI over the period of 1985-2011 was not related to biodiversity measured in 2011-2012; i.e. no effect of interannual landscape vegetation dynamics on biodiversity was found.

Additional Metadata
Keywords Agriculture, Biodiversity, Landscape metrics, Multi-spatial, Multi-temporal scale remote sensing, Vegetation phenology
Persistent URL dx.doi.org/10.1109/IGARSS.2018.8517852
Conference 38th Annual IEEE International Geoscience and Remote Sensing Symposium, IGARSS 2018
Citation
Alavi, N. (Niloofar), & King, D. (2018). Relationships of phenological and inter-annual landscape dynamics with biodiversity in farmlands. In International Geoscience and Remote Sensing Symposium (IGARSS) (pp. 5944–5947). doi:10.1109/IGARSS.2018.8517852