Farzaneh Derakhshan-Babaei sat in her Tehran office, staring at satellite images of a once-fertile catchment now scarred by patches of degradation. For years, she had watched as erratic rainfall and prolonged droughts reshaped landscapes, leaving behind a patchwork of vulnerable soils. Her latest research, published in *Progress in Disaster Science* (formerly *Pazhouhesh-e Zamin-e Sazman*), suggests that understanding these shifts isn’t just an academic exercise—it could redefine how industries like energy plan for the future.
Derakhshan-Babaei, a physical geographer at Shahid Beheshti University, led a team that developed a geospatial framework using remote sensing and machine learning to map land degradation under varying hydro-climatic conditions. By analyzing data from 1994 and projecting trends to 2024, the study uncovered surprising shifts in degradation patterns. “What surprised us most was how drastically the risk landscape changed,” she says. “In wetter periods, degradation became more fragmented, almost like a mosaic of hotspots. But in dry conditions, severe degradation barely budged—remaining stubbornly persistent.”
The research used Random Forest and Support Vector Machine classifiers to assess degradation risk, with Random Forest outperforming SVM in predictive accuracy (R² = 0.54–0.68). Key drivers? Soil moisture, the aridity index, and precipitation stood out as the dominant controls. The team found that high and very high degradation areas shrank from 24% of the catchment in 1994 to just 12% by 2024, a shift that might suggest improving conditions—or at least better land management. Yet beneath that headline, a more troubling trend emerged: severe degradation in dry zones remained nearly unchanged, holding steady at 5–6%. “In water-limited environments, the damage is lasting,” Derakhshan-Babaei notes. “Recovery is slow, and the system is less resilient.”
For industries like energy—particularly those reliant on land stability for infrastructure, water access, or biofuel feedstocks—the implications are significant. The study highlights how hydro-climatic variability can accelerate or stall degradation, directly affecting operational risk. The southern catchment, for instance, consistently showed the highest degradation risk due to urban sprawl, agriculture, and geological fragility. “If you’re planning a solar farm or a biomass facility in that region,” Derakhshan-Babaei warns, “you can’t ignore the long-term stability of the land. Erosion, soil compaction, and reduced water retention could undermine your investment.”
The research also underscores the value of integrating machine learning with environmental datasets. Traditional degradation models often struggle to capture the nonlinear relationships between climate, soil, and land use. By combining remote sensing with advanced classifiers, the team created a tool that doesn’t just diagnose past damage—it forecasts future vulnerability. “This isn’t just about mapping,” she says. “It’s about anticipating where the next pressure points will be.”
Looking ahead, the framework could be scaled to other regions, offering energy companies and policymakers a way to align land use with climate realities. For sectors dependent on land resilience—whether for mining, agriculture, or renewable energy—the study’s message is clear: hydro-climatic shifts aren’t just background noise. They’re a critical variable in long-term planning. And as Derakhshan-Babaei’s work shows, ignoring them could come at a cost.