Ekbal Hussain

Vegetation change in the Sahel has long been framed through competing narratives on desertification and greening, but existing literature often overlooks sub-annual patterns. Therefore, this research investigates the timing, duration and magnitude of vegetation growth in the West African Sahel, which is important for climate-sensitive livelihood practices such as pastoralism and semi-permanent agriculture. Our research shows that studying the range between the seasonal maximum and minimum vegetation conditions (further defined as ‘seasonal envelope’) is more informative than looking at the annual means: it reveals which part of the seasonal cycle is changing and why that matters functionally. Additionally, linear regression modelling indicates that, while climate variables explain most of the seasonal variability, consistent prediction errors at the driest and hottest extremes point to non-linear vegetation responses.Existing studies often assess vegetation dynamics using long-term satellite-derived Normalized Difference Vegetation Index (NDVI) records. Following this approach, our study uses combined Landsat and MODIS datasets to examine vegetation dynamics over the last 45 years at 500 x 500m resolution. Rather than focusing on annual-average trends, wet-season peak productivity and dry-season minimum conditions are analysed as seasonal indicators that directly guide land use, resource access, and mobility decision of local inhabitants, including pastoralists and agriculturalists in Ghana, Mali and Nigeria.Over recent decades, dry-season NDVI minima have remained relatively stable, while wet-season NDVI maxima have flattened or declined across large parts of the study regions. This divergence suggests a contraction of the seasonal NDVI envelope driven primarily by reduced peak productivity rather than declining baseline vegetation conditions, as often suggested in literature. Time-series decomposition of month-to-month NDVI variability, combined with analysis of long-term seasonal means and seasonal peak values reveals functional asymmetries in vegetation response over the last 45 years.Climate variables explain most of the interannual variability and long-term trends in linear predictive NDVI models, yet systematic modelling errors at seasonal extremes indicate the influence of non-climatic factors. To distinguish where vegetation dynamics are climate-driven or not, the second part of the study moves from historical analysis to NDVI modelling using rainfall, temperature, soil moisture, and lagged NDVI as predictor. A linear regression framework is applied, not to maximise predictive accuracy, but to diagnose where standard climate-NDVI relationships succeed or fail in capturing month-to-month vegetation responses. The largest deviations are found under high near-surface temperatures and prolonged low-precipitation conditions, proving that vegetation responses in very hot, dry circumstances are not fully captured by linear climate-NDVI relationships. These nonlinear responses are particularly noticeable during the dry-season extremes.These results establish a baseline for interpreting vegetation change that is directly relevant to land-use monitoring, early-warning systems, food security planning, and climate adaptation policy. In semi-arid regions, where livelihoods depend on narrow windows of resource availability, small shifts in wet-season productivity or dry-season duration can result in large socio-ecological consequences.

The Bandung Metropolitan Region in West Java, Indonesia, is home to over 9 million people and is exposed to multiple interacting geological and hydrometeorological hazards. Bandung City is located within a basin bounded by the Lembang Fault and the active Tangkuban Perahu volcano to the north. Previous studies indicate that the Lembang Fault is capable of generating an earthquake of up to magnitude 7, potentially subjecting up to one third of the region to severe ground shaking. Parts of the city are also highly susceptible to landslides, which, when combined with seismic activity, create the potential for cascading hazards with compounding impacts. In addition, Bandung frequently experiences flooding and landslides that have displaced tens of thousands of people. The city is further exposed to volcanic hazards from nearby volcanoes, including Tangkuban Perahu to the north and Guntur, Papandayan, and Galunggung to the southeast.This research explores both single- and multi-hazard events that could impact Bandung City using a combination of qualitative and quantitative approaches. Co-designed with the Bandung City Government, the study responds directly to the need for improved understanding of the city’s hazard landscape and patterns of exposure to inform effective risk reduction and resilience-building interventions. We employ a storyline approach to develop plausible multi-hazard scenarios involving earthquakes, volcanic eruptions, and associated cascading hazards. These scenarios are complemented by quantitative modelling of earthquake and volcanic hazards, including tephra fall, lahars, and pyroclastic density currents, to produce probabilistic hazard footprints for disaster risk management planning. In parallel, we are developing detailed physical and social exposure models using satellite Earth observation, artificial intelligence, and census data integration to identify communities most at risk. Through close collaboration with local authorities, we are engaging directly with these communities via workshops to better understand vulnerability and co-develop targeted interventions that enhance community resilience and support sustainable development by reducing long-term human and economic losses.

In the first few hours of responding to natural hazards it is crucial to understand the size of the hazard event and the scale of the potential humanitarian emergency. This is important for the timely activation of appropriate aid and support mechanisms. For earthquakes the most reliable source of immediate scientific information is from the United States Geological Survey (USGS). Through their PAGER system the USGS provide rough estimates of potential fatalities and economic impact of major earthquake events (Jaiswal et al., 2010). However, these estimates lack spatial granularity in addition to age and gender disaggregation. We know that children, elderly and women are more prone to negative impacts in a disaster (e.g. Neumayer & Plümper 2007). Therefore, it is important to have a sense of these numbers as soon as possible to understand the potential scale of the emergency.Additionally, a map of the potentially affected areas is important to understand the spatial distribution of the potential humanitarian need (e.g. isolated communities, road connectivity etc.). For example, following the 2015 Nepal earthquake the immediate acute needs of remote communities of western Nepal were initially overlooked. These communities faced severe isolation due to destroyed infrastructure, making aid delivery and access to basic supplies like food, water, and shelter challenging (The Asia Foundation, 2015). Mapping the potential exposed populations and their spatial distribution rapidly can help target appropriate emergency interventions sooner.Here we present a tool developed by the British Geological Survey that automatically and in real-time extracts earthquake shaking information from the USGS and extracts statistics of the populations, disaggregated by age and gender, who will have been exposed to certain levels of shaking. We test how rapid exposures estimates, within 3 hours of an event, can capture the final losses in major earthquakes by using the 2023 Türkiye earthquakes as a case study.We also demonstrate how we can estimate exposures and potentially compounding impacts of multiple hazards on populations following major earthquakes.

As one of the most destructive natural hazards, landslides pose persistent threats to human life, property, and critical infrastructure in Indonesia, where intense rainfall and steep, complex terrain strongly control landslide occurrence and impacts. Although landslides may be triggered by multiple factors, including earthquakes and prolonged rainfall, rainfall remains the only trigger that can be forecasted, making it central to operational landslide early warning. Between 2019 and 2024, based on Indonesian Disaster Information Database (DIBI–BNPB), more than 4,000 landslides were recorded across Indonesia, causing substantial loss of life and widespread damage to housing and public infrastructure.At present, landslide early warning in Indonesia relies on a single nationwide rainfall threshold, which may limit forecast accuracy and reliability given the country’s strong spatial variability in rainfall patterns and geomorphological conditions. Developing rainfall thresholds at large spatial scales is therefore challenging. To address this limitation, this study adopts a zoning approach that prioritises areas with high landslide susceptibility and potentially severe impacts, providing a targeted basis for subsequent threshold development.Landslide susceptibility maps are produced using the Analytical Hierarchy Process (AHP), chosen in preference to data-driven methods due to biases and incompleteness in the available landslide inventory, which tends to reflect population distribution rather than true landslide source areas. Two provinces, Central Java and South Sulawesi, are selected as initial case studies. According to the data from Local Indonesian Disaster Management (BPBD), more than 2,000 landslides were recorded in Central Java between 2016 and 2025, while over 500 events were documented in South Sulawesi between 2021 and 2025.Population density, building distribution, landslide susceptibility, and landslide runout probability are integrated to identify zones with the highest potential impacts. These high-impact zones serve as priority areas for developing more representative rainfall thresholds, with the aim of improving landslide forecasting and risk reduction in Indonesia.

The Palu-Koro and Matano faults are among the fastest-slipping and most seismically active fault systems in eastern Indonesia, yet their kinematics and influence on landscape evolution remain debated. While topographic metrics are widely used to infer vertical displacement in extensional or compressional settings, their application in strike-slip systems, where deformation is predominantly horizontal, is less established. Here, we quantify topographic metrics including relief, channel steepness index (ksn), mean slope (S), and hilltop curvature (CHT) along both faults to characterise landscape response and identify interactions between fault segments. Using a quantile-based statistical approach, we classify topographic signals indicative of tectonic activity. Along the Matano fault, elevated metrics coincide with lithological contrasts and changes in fault geometry, whereas most of the fault exhibits subtle strike-slip-dominated topography. Along the Palu-Koro fault, segments of pure strike-slip motion show subdued metrics, while areas of complex geometry and transtension display elevated values. Subdued topographic segments also spatially correspond to zones of seismic quiescence. Preliminary InSAR observations suggest creeping behaviour in the western Matano fault and the aseismic portion of the Palu-Koro fault. These findings indicate that, although topographic metrics may not directly diagnose frictional slip modes, they effectively map kinematic segmentation and structural complexity that control vertical deformation in strike-slip systems. Integrating topographic metrics with geodetic data provides a powerful approach to identify and understand fault segmentation and interaction in complex strike-slip environments.

Rapid urbanization and groundwater extraction in Java’s coastal cities are driving land subsidence at rates that far exceed global sea-level rise, amplifying flood risk and coastal vulnerability. Using Sentinel‑1 data from 2016 to 2023, we generated an InSAR-derived vertical land motion map for the island and integrated it with geology, land cover, elevation, and socioeconomic datasets. The InSAR time series was processed using a Small Baseline Subset approach and calibrated against 59 GNSS stations. Subsidence patterns were analysed through spatial correlation and Random Forest regression to identify relationships between ground motions and lowland alluvial plains in the northern part of the island. Results show that 10% of urban areas are subsiding faster than 14 mm/yr, and 5% exceed 28 mm/yr, with hotspots such as Pekalongan and Demak reaching city-wide median subsidence rates of 87 mm/yr and 43 mm/yr, respectively. Subsidence is most pronounced in the low‑lying northern plains, where the shallow geology is dominated by unconsolidated alluvium and clay-rich deposits. Currently, at least 7 million people and 920,000 buildings are exposed to subsidence ≥14 mm/yr in coastal Java. Projections under high-emission scenarios (RCP 8.5) indicate that by 2100, up to 1.4 million residents could live at or below sea level when land subsidence is combined with climate change-driven sea-level rise. Our findings underscore the need for locally tailored and integrated adaptation strategies that include urban development planning, groundwater regulation, and coastal defences, to mitigate the compounding hazards from subsidence and climate-driven sea-level change.

Coseismic Coulomb Stress Transfer (CST) from normal faulting earthquakes generally promotes failure of along-strike faults or segments, whereas it reduces stress on across-strike faults. Nevertheless, large earthquakes occur on across-strike faults despite receiving repeated negative coseismic CST. This study explores the triggering mechanisms of large earthquakes in the rapidly deforming, geometrically complex normal fault network of SW Türkiye. We model 400 years of cumulative coseismic CST from 26 M≥6.0 earthquakes plus interseismic CST based on fault slip rates. Subsequently, uncertainties in slip rates (therefore interseismic loading) and locations of historical earthquakes are explored. We find that interseismic loading is the dominant triggering process, (over-)compensating negative coseismic CST. Even CST from repeated large earthquakes on across-strike fault zones will only delay failure but is insufficient to completely suppress large earthquakes. Relatively quiescent faults in such settings should therefore not be disregarded as hazardous seismogenic sources. Nevertheless, in geometrically complex settings, heterogeneous coseismic CST might form stress barriers, reducing rupture extent and magnitude of subsequent earthquakes. Incorporating interseismic loading into Coulomb stress models is vital for robust interpretations of fault interactions, even if fault slip rates used to model interseismic loading contain major uncertainties, as accuracy of fault slip rates used is of secondary importance. Finally, the model allows interpretations of the current stress state of the regional fault network, though interpretations regarding seismic hazard should consider the described model limitations. We emphasize the need to assess uncertainties in Coulomb stress modelling studies, especially when using the historical earthquake record.