Conference Agenda

The agricultural sector is the biggest and least efficient water user, accounting for around 80% of total water use in Northern Africa, which is already strongly impacted by climate change with prolonged drought periods, imposing limitation to irrigation water availability. The objective of this study was to develop a procedure for the monitoring and forecasting of anthropogenic irrigation water use for the irrigation districts of Doukkala in Morocco from 2013 to 2022.

The analysis is based on the FEST-EWB model, that computes continuously in time both soil moisture (SM) and evapotranspiration based on the energy-water balances. The model has been calibrated and validated over non-irrigated areas, against land surface temperature (LST) from downscaled Sentinel3 data at 30m and modeled evapotranspiration from MOD16, GLEAM and FAOWapor. The model has been run using as input the past meteorological forcings (ECMWF ERA5-Land) and vegetation data from Sentinel2. Then, the actual irrigation volumes have been estimated through the calibrated model implementing three different irrigation strategy: the FAO approach based on SM crop stress thresholds (Allen et al., 1998), the separate and joint assimilation of satellite LST and SM (1km SMAP-Sentinel1) to update the modeled fluxes and estimate the irrigation volumes. Overall, the results suggested that the yearly total irrigation volumes modeled with the FAO approach are quite in agreement with the observed water allocations; and similar outcomes are obtained when the joint assimilation of satellite LST and SM. Finally, the calibrated model was used to implement a seasonal forecasted optimized irrigation strategy, by implementing the FEST-EWB model with seasonal meteorological forecast (ECMWF data) and the FAO irrigation strategy to provide forecasted crop water demand and optimized irrigation water needs.

This research was developed within the ESA AFRI-SMART project EO-Africa multi-scale smart agricultural water management, in the framework of EO AFRICA activities as Natioanl Incubators.

The anticipated increase in agricultural water usage, driven by a warming climate and rising population, underscores the need for operational large-scale monitoring tools in agricultural water governance. In this context, irrigation mapping plays a crucial role in sustainable management of agricultural resources.

Remote sensing technology offers a cost-effective alternative to traditional census methods for monitoring agricultural water usage. It provides spatially exhaustive near-real-time data, enabling timely interventions and adaptive management strategies.

In the context of the ESA Sentinel for Agricultural Statistics (Sen4Stat) project, a pixel-based map of irrigated areas was generated in Spain for 2022, using Sentinel-2 image time series, as well as high-quality in-situ datasets provided by the Spanish National Statistical Office and farmers' declarations.

Specific statistical and temporal metrics were designed to highlight phenological differences between irrigated and rainfed parcels at specific dates and throughout the growing season. To account for varying phenological response to irrigation across crop types, we employed a pixel-based categorical gradient boosting model. This model classified each pixel with known crop types (information from the farmers' declarations) as rainfed or irrigated. The overall accuracy of the obtained map is 91%; this good performance is also observed for individual crop types across a range of agricultural environments, spanning from olive groves in Andalucia to intensive cereal fields like barley in Castilla y Leon, and citrus orchards lining the Mediterranean coastline.

In the Sen4Stat context, this map will be used to update the statistical survey sampling frame, but it can also be very useful to support the transition towards more sustainable agriculture. In the future, it is planned to test the potential of the developed method in other places of the world.

Irrigation is the heaviest human alteration on the natural water cycle. Satellite technology represents a unique tool for disclosing irrigation dynamics (i.e., extent, timing, and quantification) that are generally poorly documented despite their primary importance in water resources management. In this contribution, we retrace main findings obtained in the fields of irrigation mapping and quantification through remote sensing data under the ESA (European Space Agency) funded Irrigation+ and 4DMED-Hydrology projects. The achieved results consist in the development of an irrigation mapping method called TSIMAP (Temporal-Stability-derived Irrigation MAPping) and of a framework for estimating irrigation water use, i.e., the SM-based (Soil Moisture-based) inversion approach. The latter led to the development of the first satellite-derived, high-resolution irrigation water use data sets, produced over four Mediterranean basins (the Po basin in Italy, the Ebro basin in Spain, the Medjerda basin in Tunisia, and the Herault basin in France) and over the Murray-Darling basin in Australia. In a new precursor project funded by ESA, i.e., the CCI-AWU (Climate Change Initiative – Anthropogenic Water Use), the SM-based inversion approach will be implemented, together with other satellite-based-approaches, to produce long-term irrigation water use estimates to benefit climate-related studies.

Groundwater plays a pivotal role for drinking water supply, agriculture and ecosystems in general. In arid areas such as the Saq-Ram aquifer at the Arabian Peninsula, groundwater levels (GWL) are strongly declining. In these regions, the observed decline in GWL can be attributed to intensive agriculture characterized by excessive groundwater usage for irrigation purposes. In the face of climate change, agricultural irrigation demands are increasing even in dry sub-humid regions in central Europe, where they previously where they previously held minor importance. As part of the KIMoDIs project, the aim of this study is to assess the effect of agriculture on the GWL using deep learning methods, focusing specifically on the federal state of Brandenburg, which is one of Germany's driest regions.

We develop an inverse approach: The predicted GWL is decomposed into partial scale-respective signals using the discrete wavelet transform. Gradient-based feature attribution methods, such as expected gradients, are being applied on these. We infer an agricultural map of the study area using optical satellite data, such as Sentinel-2, as an input and the EuroCrops dataset as a reference. That map serves as an input to a global GWL model combined with meteorological and static parameters. After model training, the computational graph of the model is expanded by the decomposed partial signals of the predicted GWL. Based on that expansion, the gradients and, hence, the agricultural attribution to the GWL is analyzed with respect to multiple scales which shall ease algorithmic sensitivity. That attribution is to be compared to crop type specific irrigation requirements. Our approach will help both groundwater management and farmers to decide on which crops, and potentially associated irrigation practices, may lead to groundwater depletion under certain climatic conditions.

Over the past 6 years Morocco has been heavily affected by droughts, with declining rainfall resulting in limited water available in reservoirs. The challenge is to optimise the water allocation between agriculture, urban areas and tourism.

This presentation will demonstrate how we use satellite data to monitor agricultural water use from basin to field level in the Tensift Basin of Morocco. Complex geospatial data of FAO’s open-access WaPOR database together with eLEAF’s high resolution data is translated into credible, tangible information that can be directly used for reporting, decision making and planning.

eLEAF and RESING co-developed the Water Consumption Dashboard (WCD) with (1) the water board (ABHT) mandated for integrated water resources management and (2) the irrigation office (ORMVAH) managing irrigation schemes and training farmers in optimal water use. The information provided is intended to be easy-to-digest and aimed at professionals without prior GIS experience. The dashboard provides users with actual information every ten days.

Furthermore for the Haouz plain we assessed the conversion to drip irrigation in 2018. Long-term WaPOR analyses indicate a slightly decreasing trend in water consumption (ETa) of -4,2%, and increasing trends agricultural productivity (NPP +3,5%) and water productivity (WP +6,7%). The analysis does also reveal a reduction in cultivated areas. However, because of the recent high interannual fluctuations of precipitation and the stop of surface water allocation in 2021 from reservoirs and the limited series of data available after conversion, does not allow definitive results on the impact of modernization (yet).

The cooperation with ABHT and ORMVAH recently entered a two years demonstration phase to internally adapt the WCD to identify illegal groundwater withdrawals, monitoring of water consumption, crop area and development and irrigation efficiency. The developed tools contribute to more transparent, equitable and sustainable water use.

African drylands are particularly important for research and innovation because they support 75% of the continent’s agriculture; are highly vulnerable to the impacts of climate change; have high rates of food and economic insecurity that trigger humanitarian crises; are biodiversity hotspots; and play an important role in regulating the Earth's climate system via carbon storage and rainfall triggered by evapotranspiration (i.e., “rainbow water”). Researchers and advisory services increasingly leverage advanced remote sensing technology such as the ECOsystem Spaceborne Thermal Radiometer Experiment on Space Station (ECOSTRESS) and evapotranspiration (ET)-derived metrics like the evaporative stress index (ESI) to monitor agricultural droughts in African drylands. However, there is an opportunity to enhance the reliability of these products with the emergence of a new generation of hyperspectral remote sensing missions: Hyperspectral Precursor and Application Mission (PRISMA) and the Environmental Mapping and Analysis Program (ENMAP). In this study, we employ eddy covariance flux towers in Kapiti Ranch of Machakos County, Kenya to evaluate the performance of ECOSTRESS ET and ESI products driven by PRISMA and ENMAP like hyperspectral vegetation indices resampled from fluorescence box (FloX) ground spectra. The enhanced products are demonstrated with actual PRISMA and ENMAP imagery. Our findings reveal the superiority of hyperspectral narrowbands over thermal infrared, particularly in distinguishing transpiration and soil evaporation dynamics. Such information will feed into a potentials and limitations analysis for the upcoming Copernicus Hyperspectral Imaging Mission for the Environment (CHIME) and Land Surface Temperature Monitoring (LSTM) missions. CHIME and LSTM afford the opportunity for delivering the highest quality ET and ESI estimates consistently across Africa at frequent intervals.

.