The application of the FAO WaPOR data portal to monitor efficient water use in agriculture

Abstract

Water scarcity has been a growing problem for many places around the world as water usage has been increasing with double the rate of population growth in the twentieth century. As agriculture accounts for 70 percent of global freshwater withdrawals, fresh water availability will thus face even greater stress (World Bank, 2013). Sustainable Development Goals (SDG’s) 2, 6 and 7 (UN, 2015) show that the achievement of water, food, and energy security have been set to a high priority on the global world agenda. However, achieving these goals cannot be done without a proper understanding of the interlinkages between the sectors. As water resources become more stretched, the energy and food sectors’ dependence on water implies that decision-makers in all three domains should increasingly focus on water resource management as part of their policy and practice (“Water, Food and Energy | UN-Water,” n.d.).

In light of this, the Food and Agricultural Organization (FAO) of the United Nations launched the so-called Water Productivity Open-access portal (WaPOR). The portal provides free and open access to processed satellite data that enables monitoring of land and water productivity throughout Africa and the Middle East in near real time. Crop Water Productivity is defined as the crop yield per unit of water consumed, expressed in kg/m3. The objective of this thesis is to explore and assess the available datasets provided by WaPOR to improve current water resource management practices in agriculture in the Eastern Nile Basin countries. The study focuses on the quantification of monthly water withdrawals for irrigation purposes, as well as benchmarking physical water productivity of the main irrigated crops within so-called Agro-Ecological Zones of each country. Throughout this study, crop water productivity is assessed and defined as the amount of agricultural yield that can be attained per unit of water that was allocated for its production, expressed in kg/m3. Data analysis and modeling are the major tools applied to assess spatial variation of water withdrawals and water productivity and subsequently to explain the results. The results are both the quantification of monthly water withdrawals for irrigation purposes, as well as benchmarking crop water productivity of the main irrigated crops within of the countries of the case study: Egypt, Sudan and Ethiopia.

Irrigation is considered the largest water-consuming sector in the world and has great potential to become more water-efficient. Rain-fed agriculture, however, does not influence the water balance within a catchment and can thus not improve its water efficiency. Separating irrigated agriculture from rain-fed agriculture can be done by splitting the total evaporation in so-called green and blue water evaporation. Evaporation from green water is the part of the actual evaporation that is derived from rainfall that infiltrated into the soil, while evaporation from blue water is due to the use of human-made infrastructure such as pumps, with the purpose of irrigation. With blue water evaporation, the total water consumption [m3] that was used for irrigation can subsequently be calculated. This can be used to then calculate the water productivity, but also provides insight into the current water management practices of a country. The principle of the Budyko Curve has been applied to obtain blue water evaporation (Budyko, 1974), in compliance with the Water Accounting Plus procedure that was developed at IHE Delft by Wim Bastiaanssen et al. (Bastiaanssen, W.G.M., Coerver, 2017). Finally, water consumption was obtained by multiplying the pixel size with the sum of the monthly blue evaporation. The outcome of these calculations provides a water consumption expressed in m3/month.

Water productivity is calculated by dividing agricultural yield [kg] by the amount of water that was consumed for its production [m3]. Agricultural yield was obtained by multiplying above ground biomass production (AGBP) with a crop harvest index according to the crop that was identified with the phenology data. To determine the specific growing season of a pixel, so-called “Start Of Season” (SOS) and “End OF Season” (EOS) phenology data are combined. By comparing the growing season of the pixel with literature from the FAO crop calendar, the crop type could be determined. The total water consumption between the SOS and EOS dekad numbers is summed to provide the total water consumption during the growing season of the crop. This way Crop Water Productivity is eventually obtained.

Throughout this thesis, the assumption was made that crops could be distinguished and recognized, based on the available phenology data. Considering the fact that a ‘no season’ label is applied when no growing season can be distinguished, agricultural cropland was thought to be identified through this method. However, from the fact that reasonable results complying with the literature are found with the use of the FAO LCC Land Cover Map, it follows that the identification of crops through phenology and blue evaporation data does not provide accurate results. This is especially the case for Ethiopia and to a lesser extent Sudan, likely due to the fact that Egypt has hardly any rainfall and therefore consists almost solely of irrigated agriculture. Similarly, ground truthing should therefore be done regarding crop identification and the presence of irrigation per pixel.

Calculating water withdrawals gave a promising outcome for Egypt, as the calculated water withdrawals were almost similar to the water withdrawals stated by AQUASTAT. However, numbers differed by a factor 10 for both Sudan and Ethiopia. When the FAO WaPOR LCC mask is applied, better results are achieved. The calculated water withdrawals for Egypt, Sudan, and Ethiopia are lower than FAO AQUASTAT’s numbers. It should, however, be noted that FAO AQUASTAT’s numbers are based on the required water withdrawals, while WaPOR calculates the effective water withdrawals. Lower values could imply low efficiencies of the irrigation systems, which is not uncommon for all three countries. With a typical irrigation efficiency of 60 to 70% (Howell, 2003), the total amount of water withdrawals can be computed with Q_irrigation/0.65 (Kwast et al., 2016). When this is taken into consideration, the results seem promising.

Overlapping phenologies of crops and the indistinct connection with the above ground biomass data are factors that caused unreliable results for the calculation of crop water productivity. The high CWP values that were found in Ethiopia are high compared to the reasonable values found in Egypt and Sudan. This could possibly be due to the fact that the pixels are wrongly identified as irrigated pixels. After all, CWP was assessed for all pixels that contain blue evaporation and was not masked with the FAO WaPOR LCC mask. It is therefore recommended to use an accurate land use mask when CWP is assessed.