Scaling effects on spring phenology detections from MODIS data at multiple spatial resolutions over the contiguous United States

摘要

Land surface phenology (LSP) has been widely retrieved from satellite data at multiple spatial resolutions, but the spatial scaling effects on LSP detection are poorly understood. In this study, we collected enhanced vegetation index (EVI, 250 m) from collection 6 MOD13Q1 product over the contiguous United States (CONUS) in 2007 and 2008, and generated a set of multiple spatial resolution EVI data by resampling 250 m to 2 x 250 m and 3 x 250 m, 4 x 250 m,..., 35 x 250 m. These EVI time series were then used to detect the start of spring season (SOS) at various spatial resolutions. Further the SOS variation across scales was examined at each coarse resolution grid (35 x 250 m approximate to 8 km, refer to as reference grid) and ecoregion. Finally, the SOS scaling effects were associated with landscape fragment, proportion of primary land cover type, and spatial variability of seasonal greenness variation within each reference grid. The results revealed the influences of satellite spatial resolutions on SOS retrievals and the related impact factors. Specifically, SOS significantly varied lineally or logarithmically across scales although the relationship could be either positive or negative. The overall SOS values averaged from spatial resolutions between 250 m and 35 x 250 m at large ecosystem regions were generally similar with a difference less than 5 days, while the SOS values within the reference grid could differ greatly in some local areas. Moreover, the standard deviation of SOS across scales in the reference grid was less than 5 days in more than 70% of area over the CONUS, which was smaller in northeastern than in southern and western regions. The SOS scaling effect was significantly associated with heterogeneity of vegetation properties characterized using land landscape fragment, proportion of primary land cover type, and spatial variability of seasonal greenness variation, but the latter was the most important impact factor. (C) 2017 International Society for Photogrammetry and Remote Sensing, Inc. (ISPRS). Published by Elsevier B.V. All rights reserved.