Unmasking Nile’s Land Grabbed Territories
The advanced political economies have shape a world where complexity too often tends to produce elementary brutalities under complex modes of expulsion. Most of the time those dynamics operate at a more subterranean level. The tools driving this phenomenon range from basic policies to complex institutions, systems, and techniques that demand specialized knowledge and sophisticated organizational structures.
The dynamics of expulsion can take different shapes, in our proposal we focus on the complexity of the legal contracts, that enabling a sovereign country to acquire vast stretchers of land in a foreign sovereign nation- state as a source of extension of its own territory, even as it expels local villages and rural economies from that land. (Sassen, 2014).
Furthermore the advanced economic and technical achievements enable the extraction of natural resources, such as water and land, destroying the environment. The availability of freshwater is becoming a global concern. Because agricultural consumption has been increasing steadily, the mapping of irrigated areas is key for supporting the monitoring of land use and better management of available water resources.
The project propose a detection method to identify center pivot irrigation systems by using ONNX Neural Network models, along Nile River water basin.
Our goal is to provide a fast and accurate alternative to map center pivot irrigation systems for future social and environmental analysis. This method, has the potential to be scaled to larger areas for identify these irrigations systems along the world, use by agricultural activities.
Land grab
“Transnational land acquisitions refer to the procedure of acquiring land (and freshwater) resources in foreign countries. It is often called ‘land grabbing’. Most commonly, investors or investing countries are located in the developed world, while the ‘grabbed’ land is usually in developing countries.”
European Environment Agency
Water grab
“A situation where powerful actors are able to take control of, or reallocate to their own benefits, water resources already used by local communities or feeding aquatic ecosystems on which their livelihoods are based”
Mehta, S (2012). Introduction to the Special Issue: Water Grabbing? Focus on the (Re)appropriation of Finite Water Resources. Water Alternatives.
Worldwide scale
Mapping the land grab
We use Land Matrix , and EU, NGO and Academic collaboration for investigate mass land acquisition with traits of land grabbing. These dataset allows for a georeferenced investigation of land acquisition to create a network of worldwide transactions.
In the map on the right, we can observe the five top buyers countries and the selling countries, mostly countries in the global south.
Within the five top countries, we found Cyprus as one the major buyer. This is because the country is a tax haven and many companies register and conduct transactions from there.
Mapping water and land grab
The image above (left), show the different location of each deals, between the company countries and the target countries, based on Land Matrix. Each of its deals have specific information about the agreements, such as: amount of water, type of crop, irrigation system used, amount other.
In the other hand, the image on the right side represent the river basins areas and the different levels of water stress based.
If we overlap those layers, we can observe a bigger correlation between the high level stress water and the high concentration of land acquisition deals. In order to take a deep look to this, we determinate a region of interest, with special attention to the territories along the Nile River.
Region of interest: The Blue Nile
Originating from Lake Tana in Ethiopia, supplying about 85% of the Nile’s total flow. Beyond agriculture, the river plays a crucial role in hydropower, with Ethiopia’s Grand Ethiopian Renaissance Dam (GERD) with the aim to boost energy production.
However, this has sparked geopolitical tensions over water access, as any changes in flow could impact farming and food security downstream.
See the Nile beyond political borders
Because water knows no borders, it is important to look at the Nile as a whole territory and aknowledge that there are different laws and systems that have an impact on the whole river. Water extraction in Ethiopia directly affects Egypt because 80% of the Nile River’s water comes from Ethiopia.
Any alteration in the flow of the Blue Nile impacts the water supply of Egypt, a country that relies almost entirely on the Nile for its agriculture, industry and human consumption.
How can remote sensing techniques, employing machine learning-based object detection, reveal extractive corporate practices in the Nile Basin?
Center-pivot irrigation systems serve as a readily detectable feature and a visual proxy for identifying extractive agricultural practices.
[…] a self-propelled apparatus for sprinkling and irrigating […] large sections of land…
In sections of the country where the natural rainfall is not sufficient […] it is sometime absolutely necessary to provide additional water.
This water may be taken from a river or a lake, perhaps some distance away.
Or the water may be obtained from an artesian well or other underground source.
High Water Consumption – Extracts large amounts of water, leading to depletion of groundwater and surface water sources.
Salinization – In arid regions, repeated irrigation can lead to salt buildup in the soil, reducing crop productivity.
Soil Degradation – Over-irrigation can cause soil erosion, compaction, and nutrient leaching.
Project’s Methodology
The project leverages machine learning and satellite imagery to detect pivot irrigation systems in the Nile Basin, following a structured methodology that integrates data collection, model training, and geospatial analysis.
The process begins with model creation, where 809 satellite images are collected using Python and Leafmap, then annotated semi-automatically in Roboflow. A detection model is trained over 50 epochs to recognize pivot irrigation structures. Once the model is established, large-scale data collection is conducted by defining key regions of interest in QGIS and acquiring a decade’s worth of Sentinel imagery using Python and Geemap. These images are tiled annually with Rasterio to facilitate further processing.
In the inference stage, the detection model is tested in QGIS Deepness, running object detection on Sentinel images. The results are fine-tuned through iterative testing. This leads to the refinement phase, where Python scripts filter out large, irrelevant objects and remove isolated detections, ensuring a more precise dataset.
The final stage of this pipeline is analysis, where Python quantifies detections per grid cell, and QGIS visualizes hotspots of irrigation activity. This structured methodology provides a foundation for further investigation into agricultural expansion and its environmental implications, setting the stage for broader spatial and ecological assessments.
Analysis Results
Our investigation over the decade from 2014 to 2024 reveals a gradual expansion of center-pivot irrigation systems along the Nile. During this period, the number of successfully detected objects increased fivefold. Notably, the rate of expansion accelerated significantly in the final year of the study, indicating a marked intensification of these extractive agricultural practices. The growth is particularly associated with large-scale projects, such as the one at Toshka Lakes. The cultivated areas in this region span approximately 200 × 150 km. Notably, the expansion of these farms is closely linked to the development of infrastructure, including dams, canals, and new urban settlements.
