Location: Hydrology and Remote Sensing LaboratoryTitle: A within-season approach for detecting early growth stages in corn and soybean using high temporal and spatial resolution imagery
|KARNIELI, A. - Ben Gurion University Of Negev|
|HIVELY, W.D. - US Department Of Agriculture (USDA)|
|Kustas, William - Bill|
Submitted to: Remote Sensing of Environment
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 3/5/2020
Publication Date: 6/1/2020
Citation: Gao, F.N., Anderson, M.C., Daughtry, C.S., Karnieli, A., Hively, W., Kustas, W.P. 2020. A within-season approach for detecting early growth stages in corn and soybean using high temporal and spatial resolution imagery. Remote Sensing of Environment. https://doi.org/10.1016/j.rse.2020.111752.
Interpretive Summary: Crop emergence, defined as the appearance of the first leaf, is a critical stage for crop development and crop growth modeling. Remote sensing mapping of crop emergence is very challenging because the resulting optical changes on the land surface are small. Traditional remote sensing approaches require one or more years of remote sensing data to map green-up dates. These approaches are not designed for near-real-time mapping. This paper presents a new near-real-time crop emergence (NRT-VE) mapping approach to detect crop green-up date using satellite observations during early crop growth stages. Results show that early crop growth stages are highly related to remote sensing green-up dates and can be reliably mapped two weeks after emergence. The NRT-VE approach maps early crop growth stages and provides spatial variability within the field required for crop precision management and crop yield estimation at small scales.
Technical Abstract: Crop emergence is a critical stage for crop development and crop growth modeling. Mapping crop emergence using remote sensing data is challenging. Previous remote sensing phenology algorithms showed that crop stages could be detected around the V3-V4 (3 to 4 established leaves) vegetative stage. Traditional approaches have a strong assumption regarding the temporal evolution of plant growth and normally require a complete growth period of observations to define seasonal changes. These approaches were not designed for the near real-time mapping. In the current paper, we developed a new near-real-time crop emergence (NRT-VE) approach to mapping crop green-up date using satellite observations during early growth stages. The approach was first optimized using high spatiotemporal resolution (10m, 2 day revisit) imagery from the VENµS research mission, and assessed using ground observations of early crop growth stages (VE-V1 stages for corn, VE-VC stages for soybeans) collected over the Beltsville Agricultural Research Center (BARC) experimental fields in Beltsville, MD during the 2019 growing season. Results show that early crop growth stages can be reliably detected at sub-field scale about two weeks after crop emergence. The remote sensing detcted green-up dates are about 4-5 days after emergence (VE) on average. Coefficients of determination (R2) between green-up dates and the mid-point dates of the early growth stages are above 0.90. The mean absolute differences, standard deviations, and root mean square errors comparing to the early growth stage mid-point dates are within six days. The maximum differences are within ±10 days across all fields. For independent evaluation, the NRT-VE approach was applied over an agricultural watershed on the Maryland Eastern Shore and compared with crop progress reports of emergence dates from the National Agricultural Statistics Service (NASS) at the state-level. The NRT-VE-detected green-up dates at the regional scale are within crop growth ranges but slightly earlier than the NASS crop progress reports at the state-level. The NRT-VE approach relies on remote sensing observations during the early crop growth stages and has potential for operational application in the near real-time if consistent high temporal and spatial resolution satellite data are routinely available.