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 #345042
Title: Investigating water use over the Choptank River Watershed using a multi-satellite data fusion approach

Author
item SUN, L. - US Department Of Agriculture (USDA)
item Anderson, Martha
item Gao, Feng
item HAIN, C. - Goddard Space Flight Center
item Alfieri, Joseph
item SHARIFI, AMIR - University Of Maryland
item McCarty, Gregory
item Yang, Yun
item YANG, YANG - US Department Of Agriculture (USDA)
item Kustas, William - Bill
item McKee, Lynn

Submitted to: Water Resources Research
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 10/24/2017
Publication Date: 11/3/2017
Citation: Sun, L., Anderson, M.C., Gao, F.N., Hain, C., Alfieri, J.G., Sharifi, A., McCarty, G.W., Yang, Y., Yang, Y., Kustas, W.P., Mckee, L.G. 2017. Investigating water use over the Choptank River Watershed using a multi-satellite data fusion approach. Water Resources Research. 53:5298-5319.

Interpretive Summary: The health of the Chesapeake Bay ecosystem has been declining for several decades due to high levels of nutrients and sediments largely tied to agricultural production systems within the Bay watershed. Models of the hydrologic processes within the Bay system have been developed to evaluate impacts of best management practices aimed at reducing nutrient and sediment loads to Bay tributaries. However, these models may not accurately represent the effects of increasing irrigation intensity that is occurring on agricultural lands within the watershed, particularly on the Eastern Shore of Maryland. The hydrologic models also require spatially distributed calibration information that does not currently exist with adequate spatial sampling. Additional proxy observations are required to improve these models, particularly describing changes in anthropogenic water use for irrigated agriculture within the Bay watershed. This paper explores the utility of remotely sensed evapotranspiration (ET) derived from satellite imagery, which may serve well in this capacity for proxy water use information. A data fusion method is employed to combine ET maps developed from multiple satellite data products to produce water use information at daily timescales and sub-field spatial resolution, capable of capturing individual irrigated fields. The model is evaluated using flux measurements collected at a micrometeorological tower in the Choptank River basin within the Lower Chesapeake Bay Long-term Agroecosystem Research area. The model well reproduces ET measured in an irrigated field with rotating crops. Water accounts for different landuse classes within the modeling domain were assessed. Model estimates of soil evaporation (E) and crop transpiration (T) show reasonable temporal patterns for a variety of typical single and double cropping systems. The model also demonstrates the capacity to capture enhanced water use signals over fields that are known to be irrigated. These assessments lend credibility to the use of remotely sensed ET as a proxy calibration data source for improving hydrologic models used for Chesapeake Bay water quality analyses. This calibration work is ongoing, and will be reported in a follow-up publication.

Technical Abstract: The health of the Chesapeake Bay ecosystem has been declining for several decades due to high levels of nutrients and sediments largely tied to agricultural production systems within the Bay watershed. Therefore, monitoring of crop production, agricultural water use and hydrologic connections between crop lands and Bay tributaries has received increasing attention. Satellite remote sensing technologies using thermal infrared (TIR) imaging have been widely used to map spatiotemporal variability in consumptive water use (or evapotranspiration; ET) for agricultural water management applications. However current methods based on use of single TIR satellite platforms do not provide both the temporal and spatial resolution required to meet critical application requirements – namely, daily monitoring at sub-field scales. In this study, a multi-sensor satellite data fusion methodology, combined with a multi-scale ET retrieval algorithm, was applied over the Choptank River watershed located within the Lower Chesapeake Bay region on the Eastern Shore of MD, USA to produce daily 30m resolution ET maps. The modeling system combines TIR information collected with the Geostationary Environmental Operational Satellites (GOES), the Moderate Resolution Imaging Spectroradiometer (MODIS) and the Landsat satellite suite. ET estimates directly retrieved on Landsat overpass dates have high accuracy with bias of 0.17 mm d-1, Root Mean Square Error (RMSE) of 1.18 mm, and relative error (RE) of 9%, as evaluated using eddy covariance flux tower measurements. The fused daily ET time series have reasonable errors of 18% at the daily time step - an improvement from 27% errors using standard Landsat-only interpolation techniques. An accounting of annual water consumption by different land cover types was performed, showing reasonable distributions of water use with cover class. Seasonal patterns in modeled crop transpiration and soil evaporation for dominant crop types were analyzed, and agree well with crop phenology at spatiotemporal scale. Additionally, effects of irrigation occurring during a period of rainfall shortage were captured by the fusion program. These results suggest that the ET fusion system will have utility for water management at field and regional scales over the Eastern Shore. Further efforts are underway to integrate these detailed water use datasets into hydrologic modeling to improve assessments of water quality and best management practices within the Chesapeake Bay.