Yu Shen and Xiaoyang Zhang, professors in the Department of Geography and Geospatial Sciences, South Dakota State University, Brookings, SD 57007, USA, are investigating whether it would be possible to make crop real-time monitoring of crop progress more efficient.
Numerous previous satellite observations have allowed for the widespread detection of crop phenology. On the other hand, due to a shortage of high-frequency cloud-free satellite observations and anticipated future crop development, Near-Real-Time (NRT) monitoring of crop progress using timely available remote sensing data is little explored. This paper suggests a unique algorithm for operational NRT monitoring of agricultural progress at the field scale to address the problem. This algorithm produces cloud-free time series of HLS-ABI EVI2 (two-band Enhanced Vegetation Index) with a Spatiotemporal Shape-Matching Model (SSMM) by first fusing the high spatial resolution Harmonized Landsat and Sentinel-2 (HLS) data (30 m) and the high temporal frequent (10 min) Advanced Baseline Imager (ABI) observations. Then, using a reference EVI2 time series acquired from the nearby pixels in the year before, it makes predictions about likely future EVI2 values at a specific pixel. The algorithm finally detects six crop phenometrics including greenup onset, mid-greenup phase, maturity onset, senescence onset, mid-senescence phase, and dormancy onset by integrating the currently available HLS-ABI observations and the predicted future EVI2 values to generate annual EVI2 time series. Throughout the growing season, new HLS and ABI observations are added to the NRT monitoring, which is divided into three categories: near-real-time prediction (phenological event detected after the occurrence), real-time prediction (phenological event detected around the occurrence), and short-term prediction (phenological event detected before the occurrence). We compare the results of the NRT monitoring in Iowa in 2020 to those of industry-standard phenology products, PhenoCam observations, and the weekly Crop Progress Reports (CPRs) published by the National Agricultural Statistics Service (NASS) of the United States Department of Agriculture (USDA). The evaluation shows how reliable the developed method is for NRT crop phenology monitoring. Real-time prediction reveals that the Mean Absolute Difference (MAD) is 10 days for greenup and dormancy onset, and 5 days in the other four phenometrics, even if the uncertainties are rather substantial for short-term prediction compared with routine detections. With a MAD of 7.8 days, the real-time prediction also agrees well with PhenoCam data (R2 = 0.96, P 0.001). Additionally, the HLS-ABI real-time prediction of crop phenometrics is able to closely follow NASS crop progress with minimal temporal shifts ( 5 days) and substantial correlations (R2 > 0.85, P 0.001) for various phenological phases of maize and soybean. These findings demonstrate the algorithm’s applicability for NRT monitoring of several agricultural phenometrics at multiple sizes, from the field to the national level.
This study was published in the journal Remote Sensing of Environment.