Estimating Evapotrnspiration in a Rice Field Using a Remote-Sensing Based Two Source Energy Balance Model
Abstract
Evapotranspiration monitoring of rice, a main cereal and food source of Monsoon Asia, is important not only for sustaining stable grain production and for effective water use through precise water management, but also provides a means for early warning of and response to drought. The remote-sensing based two source energy balance model (TSEB) estimates of evapotranspiration (ET) over a wide variety of land cover types using ground, airborne and satellite imagery and meteorological data without time-consuming and/or expensive field measurements such as measurements of daily decrease of flooding water depth and eddy covariance-based flux tower observations. We, therefore, evaluated the TSEB model at local sites, Icheon and Kimje, with energy flux tower and ground-based thermal-infrared temperature measurement collected over cultivated rice fields and applied the model for estimating ET over rice cropping region encompassing an area 16km x 16km scale in South Korea using Landsat imagery. The TSEB model required modification to the soil heat flux algorithm because rice typically grows in saturated soils and/or standing water for about 75% of the rice growing season. Half-hourly energy flux data, including net radiation, sensible heat, latent heat (corresponding to ET), and soil heat, at two field sites were acquired using eddy-covariance method. The root mean square difference values between predicted and observed latent heat flux ranged between 10% and 25% of the average observed latent heat flux. This is comparable to the measurement uncertainty, suggesting that the TSEB model can provide reliable ET estimation for rice fields. Applying the TSEB model with Landsat imagery over a 16km x 16km domain encompassing the Kimje flux tower site was also performed. Leaf area index for the study area at the Landsat resolution was estimated using MODIS leaf area index products as a reference. Atmospheric correction of the land surface temperature was carried out using MODTRAN. The ET estimated by the TSEB model for the rice field surrounding the Kimje flux tower was in close agreement with the flux tower measurement, similar to results using local ground-based land surface temperature data. Plans are to conduct future validation studies using satellite data under different soil moisture rice plant phenological conditions. In addition, issues involved in extrapolating meteorological forcing data over large areas and interpolation between Landsat satellite acquisitions will be discussed.
- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2013
- Bibcode:
- 2013AGUFM.H43G1566H
- Keywords:
-
- 1855 HYDROLOGY Remote sensing;
- 1880 HYDROLOGY Water management;
- 1847 HYDROLOGY Modeling;
- 1812 HYDROLOGY Drought