Evaluating the robustness of NISAR's cropland algorithm to time of observation, observing mode and dithering
Abstract
Cropland mapping is important for monitoring agricultural practices, cropland distribution and for supporting food security programs. The NISAR mission will support these efforts by producing a global cropland product on annual basis and at 1 ha spatial resolution. The scale of this product necessitates a straightforward approach that has low computational cost and is expected to yield reliable results under a range of conditions. In light of prior evaluations using L- and C-band (PALSAR, UAVSAR, AgriSAR, Sentinel), the cropland product will be based on a temporal Coefficient of Variation (CV) approach. With many of the NISAR's simulated products now available over agricultural areas, we will be able to further test the robustness of the CV-based cropland classifications. This work employs the NISAR simulated datasets to compare and contrast how cropland classifications vary depending on time of day, observing mode and the degree to which data are dithered. Specific research questions are: (1) How much do diurnal effects (and associated differences in incidence angles) impact the cropland product? (2) Are there any substantial accuracy differences between the observing modes to be used globally (129A, 20 MHz) and over the contiguous United States (138A, 40 MHz)? (3) How does the degree of dithering (none, gaps, no gaps) impact the algorithm's capability at identifying cropland?
- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2020
- Bibcode:
- 2020AGUFMB063.0003K
- Keywords:
-
- 0434 Data sets;
- BIOGEOSCIENCES;
- 0452 Instruments and techniques;
- BIOGEOSCIENCES;
- 0480 Remote sensing;
- BIOGEOSCIENCES;
- 1813 Eco-hydrology;
- HYDROLOGY