Mapping evapotranspiration with energy balance models by Dr. Hadi Jaafar

​Groundwater levels in some regions of Lebanon’s Bekaa Valley, the primary agricultural production area in the country, have dropped with population growth in the area. Dr. Hadi Jaafar, Assistant Professor in Irrigation and Water Management at FAFS, addresses agriculture water management issues and related challenges through his work on applications of remote sensing and geographic information system (GIS). Jaafar’s research focuses on mapping evapotranspiration (ET), the assessment of water uptake by the plant at the field scale using remote sensing imagery, specifically thermal imagery. His research is based on work he has conducted at FAFS, and is currently carrying out as Lebanon-PI in a three year NASA (National Aeronautics and Space Administration) funded grant. The grant supports collaboration with the Hydrology and Remote Sensing Lab of the United States Department of Agriculture (USDA) in Beltsville, Maryland. 

The major consumer of water in the world, agriculture also accounts for more than 85% of the total water diversions in Lebanon. Dr. Jaafar and his team are the first to generate a time-series of water use maps for Lebanon using energy balance models. The significance of these field scale maps lies in their ability to identify the quantity of water used by agriculture in each field and enabling answers to questions about the value of irrigation, whether it is worth irrigating a given crop, and the economic value of the water’s productivity.

Dr. Jaafar highlights the importance of generating information about the exact amount of water consumed by each crop, explaining that the multiple demands for water often conflict. The same sources of water are used for multiple purposes, in the Bekaa Valley, these demands include the transient, domestic, industrial and agricultural water sectors. The objective is to maximize the benefit from each unit of water. Through ET mapping, the team has data that allows them to advise the government and guide policies on maximizing available water productivity, particularly important for water-scarce regions like the Middle East.

Dr. Jaafar and his team collected data on crops in the Bekaa Valley using the variables of biomass, yield, and leaf area index, among others. The variables were then correlated with remotely-sensed indices, such as the normalized difference vegetation index (NDVI), local weather data, and Landsat remote sensing imagery to generate ET maps. The water analysis is coupled with yield analysis to estimate crop production in the planted area.

Validation is always a challenge, especially for large areas: the the team is working with pySEBAL and METRIC models, validated in different areas of the world. One part of the three-year project with NASA is assessing validity of the product in semi-arid areas, especially in the Bekaa Valley and to this end, during the first year of the project, significant amounts of data were collected on different crops to validate the model output. An additional major challenge is the heterogeneity of the agricultural land, i.e. bare soil and buildings, when contrasted with the US, where larger areas of agricultural lands and less construction dilute these challenges. 

Additionally, while METRIC and pySEBAL are single-source energy balance models, a third model, DisALEXI, is also used by NASA and USDA as a two-source energy balance model. Comparing these models will demonstrate which is performing best. However, the uncertainty in these models is low, especially regarding current estimates of water use in Lebanon. For example, a 15% error is considered low as it is still within the range of uncertainty of other applications: estimating the agricultural water use with a ±15% error on a daily scale, further decreases the weekly and monthly scales for data validation.

Dr. Jaafar and his team have recently published the article Time series trends of Landsat-based ET using automated calibration in METRIC and SEBAL: The Bekaa Valley, Lebanon​ in the respected journal: Remote Sensing of Environment. Their study, now completing its second year, analyzed more than 1,300 satellite images and required several days of computer runs. Based on these images, the team is now able to generate a weekly ET water use map for Lebanon that allows evaluation of how much water is being used in agriculture and in  other landscapes (forests, bare grounds, shrubs, trees, etc.). Findings were validated at the monthly scale using groundwater data collected from AREC over three years: the team overlapped the two data sets with other observations and found that the results of the model can be used to estimate groundwater depletion in the Bekaa.

Dr. Jaafar noted that the next step is to make these ET maps public and link them to a mobile app through which farmers can know how much water their land uses and irrigate accordingly. He and his team are currently applying for additional research grants to obtain the needed funds.

 ​

Click on the map for the high-resolution image.