Evaluation and generation of consistent high spatial resolution multispectral reflectance time series using Planetscope and NASA Harmonized Landsat Sentinel-2 data

Abstract:

Under NASA Commercial SmallSat Data Acquisition (CSDA) program funding we evaluated the utility of Planetscope-0 and Planetscope-1 imagery for burned area mapping and for validation of a NASA Land Cover Land Use Change funded 30 m Landsat-8 Sentinel-2 burned area product. This evaluation highlighted the utility of Planetscope imagery for land monitoring applications due to its near daily coverage and 3 m spatial resolution, although issues with the spectral and temporal consistency of the different Planetscope sensor generations and the absence of a Short-Wave Infrared (SWIR) band were highlighted. In the last decade, there has been an evolution in the development of analysis ready data (ARD) processed in a community endorsed manner, including the NASA harmonized Landsat and Sentinel-2 surface reflectance data set (HLS). We, and other researchers, have developed algorithms to blend Landsat and MODIS to generate synthetic daily Landsat 30 m data, and more recently to generate synthetic daily Landsat reflectance without using other satellite data. With the availability of commercial high resolution data there is considerable interest in refining these approaches to generate a consistent high spatial resolution ARD. We present research that (1) quantifies the spectral differences between the different generations of Planetscope sensors; (2) quantifies the spectral differences between Planetscope reflectance and NASA HLS 30 m nadir BRDF adjusted reflectance (NBAR), (3) sharpens to 3 m the 30 m HLS surface NBAR for the visible, NIR and SWIR bands. The next steps and potential solutions to generate 3 m visible to SWIR surface reflectance ARD that are consistent with the NASA HLS are discussed.