Skip to main content
ARS Home » Northeast Area » Beltsville, Maryland (BARC) » Beltsville Agricultural Research Center » Hydrology and Remote Sensing Laboratory » Research » Publications at this Location » Publication #326232
Title: Monitoring crop condition at field scale using multiple remote sensing data

Author
item Gao, Feng
item Anderson, Martha

Submitted to: Meeting Abstract
Publication Type: Abstract Only
Publication Acceptance Date: 2/15/2016
Publication Date: 7/18/2016
Citation: Gao, F.N., Anderson, M.C. 2016. Monitoring crop condition at field scale using multiple remote sensing data. Meeting Abstract. 2016 CDROM.

Interpretive Summary:

Technical Abstract: Crop growth condition is affected by both environmental variables (climate, weather and soil condition etc.) and anthropogenic activities (fertilization and irrigation etc.). Crop condition varies by year and location and is critical for crop management and yield estimation. In the United States, crop condition information is published weekly by the U.S. Department of Agriculture (USDA) National Agricultural Statistics Service (NASS). Local observers report crop growth stages and conditions according to the NASS’s standard definition. The crop progress reports are then summarized at state and district (multiple counties) level. These reports do not include spatial variability within state or district. The information is subjective to observers’ judgment and may lack of consistency among different statistical units. For precision agricultural management, monitoring crop condition at field scale is required. Satellite remote sensing provides consistent information for monitoring crop conditions. Coarse resolution data such as AVHRR and MODIS have been used to monitor crop condition. However, the 250m to 1km resolution imageries are too coarse to distinguish individual fields. In order to monitor crop conditions at field scale, high spatial and temporal resolution remote sensing data are required. However, high spatial and temporal resolution remote sensing data are currently not available from a single sensor. Data fusion approach provides a way to generate such data set from multiple remote sensing data sources. In this work, the Spatial and Temporal Adaptive Reflectance Fusion Model (STARFM) was used to generate daily Landsat-like surface reflectance. Our study area locates in a rain-fed agricultural area in central Iowa. Landsat and MODIS surface reflectance from 2001 to 2014 were ordered and processed. Using the 14 years of fused Landsat and MODIS data, a Look-Up Map (LUM) of Normalized Difference Vegetation Index (NDVI) for each major crop type in the Cropland Data Layer (CDL) was generated using daily NDVIs from the years under good conditions. Gap-filling approach was developed to replace the missing pixels in the LUM due to missing of NDVI or crop sample. Eventually, a spatially complete LUM for each crop type was built. The LUMs considered the phenology difference among different years. The LUMs include the mean and standard deviation of NDVI for each pixel under good condition starting from 30 days before the start of the season. The fused Landsat and MODIS data for years under poor conditions were examined and compared to the LUMs. The differences of NDVI were examined and compared to the NASS Crop Progress reports. Once the NDVI LUMs become stable, we will apply the approach to the near real time remote sensing data.