Atmospheric corrections of WorldView-3 imagery to improve the accuracy of space-borne estimates of crop residue cover assessments in Mid-Atlantic agricultural regions.
Abstract
Agricultural regions exhibit highly dynamic reflectance properties on monthly and even weekly timescales. Planting, harvesting, and tilling produce large changes in surface reflectance as agricultural fields transition from exposed soils or soils covered with plant litter to mature plant canopies and back. Satellite imagery enables effective monitoring of these changes on regional scales. Plant health, soil moisture, and tillage practices can all be assessed via. spectral indices and algorithms derived from multispectral satellite imagery. These indices and algorithms are most biophysically linked to agricultural field condition when surface reflectance satellite imagery is used to produce them. However, spatiotemporal heterogeneity of the atmosphere inhibits the accurate characterization of surface reflectance, and reduces the ability to accurately assess agricultural change with multi-date imagery. Ozone, NO2, CO2, and water vapor all exert strong impacts on the transmission of light between the Earth's surface and the top of the atmosphere. Atmospheric corrections are needed to control for these factors.
In this research effort, we used high spatial resolution WorldView-3 imagery to derive spectral indices for characterizing tillage practices over agricultural study sites on the Eastern Shore of Maryland. In the first section of this presentation, we discuss how MODTRAN 5.3.3. was used to process WorldView-3 visible-near infrared (VNIR) and shortwave infrared (SWIR) top of atmosphere (TOA) radiance imagery to surface reflectance imagery. We also compare results from MODTRAN to Simple Model of the Atmospheric Radiative Transfer of Sunshine (SMARTS) and Fast Line-of-sight Atmospheric Analysis of Hypercubes (FLAASH). In order to assess the accuracy of these atmospheric correction approaches, we compare their corrected surface reflectance values to proximal surface reflectance values obtained from a CropScan spectroradiometer. We also present results highlighting the degree of importance of atmospheric corrections in terms of image-to-image comparability. In the second section of this research effort, we present a WorldView-3-based crop reside cover time series that elucidates changes in tillage practices on the Eastern Shore of Maryland between 2015 and 2017.- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2018
- Bibcode:
- 2018AGUFM.B43B..05L
- Keywords:
-
- 0402 Agricultural systems;
- BIOGEOSCIENCESDE: 0416 Biogeophysics;
- BIOGEOSCIENCESDE: 0430 Computational methods and data processing;
- BIOGEOSCIENCESDE: 0480 Remote sensing;
- BIOGEOSCIENCES