Akash Koppa

Flash droughts pose severe risks to societies and ecosystems due to their sudden onset and rapid intensification. Predicting their vegetation impacts at subseasonal timescales remains difficult. Advances in hybrid physics‐AI models and weather forecasting create new opportunities to enhance the predictability of evaporative stress, a key indicator of flash drought onset. Here, we present a novel subseasonal global framework that integrates bias‐corrected, high‐resolution weather forecasts with a hybrid evaporation model to predict flash drought onset, persistence, and termination up to 10 days in advance. The framework shows high skill in persistence and termination, especially in semiarid regions, with median Critical Success Index of 0.68 and 0.56, respectively. Over 70% of flash‐drought‐affected areas in the United States and Europe achieve a Probability of Detection above 0.5 against operational monitors. Our framework supports the development of operational early warning systems to reduce flash drought risks to food and water security.

Rainfed crops account for approximately 40% of India’s food production and support 60% of its livestock. Although linked to oceanic monsoon rainfall, their productivity also depends on terrestrially-sourced rainfall, particularly in the non-monsoon season. The degree to which rainfed crops rely on moisture sourced from evaporation in upwind irrigated areas remains largely unknown. Using a combination of models and observations, we show that evaporation from upwind irrigated crops contributes 7% (mean) ± 5% (spread) of the rainfall over rainfed areas annually, rising to 15 ± 10% during the pre-monsoon months (averaged over the years 2000–2020). In the absence of this input, water stress experienced by rainfed crops can increase by 5%–10% during the crucial mid to late crop growth phases, potentially affecting yields. Our results reveal an unrecognized atmospheric link between irrigated and rainfed agriculture that is overlooked in current agricultural policies. Planning and managing these systems holistically can help strengthen regional food and water security under future climates.

Understanding water deficit and water stress is crucial, as these factors directly influence irrigation needs, crop productivity, and long-term agricultural sustainability in regions prone to drought and uneven precipitation. This study quantifies historical and future water-deficit and heat-stress dynamics for corn and soybean in the Mid-Atlantic region of the U.S. under three emission scenarios (SSP126, SSP245, and SSP585). Water deficits were assessed using a multi-temporal cumulative deficit framework that captures the effects of precipitation timing and short-term soil moisture carryover on crop water availability. Heat stress was evaluated using crop-specific temperature thresholds during sensitive growth stages, with metrics including total heat-stress days, maximum consecutive hot days, and cumulative thermal excess. Results show consistent increases in ET0 and ETc across all scenarios, with the greatest rises under SSP585, particularly in southern and central Maryland. Water deficits intensify over time, and although multi-day precipitation accumulation reduces apparent deficits, even the lowest-emission scenario projects water-deficit values above historical baselines. Heat-stress exposure also increases, with higher-emission scenarios producing more frequent, prolonged, and intense hot periods, especially in the late century. The combined trends in water deficit and heat stress indicate that, without adaptation, climate change will significantly increase irrigation demand and crop stress risk.

Since the beginning of the 21st century, Central Europe (CE) has experienced a series of dry summers with substantial socioeconomic and environmental impacts. While the influence of the North Atlantic Ocean is well-established, the role of the Pacific Ocean in modulating CE summer rainfall (CESR) is relatively unknown. Here, we present an evidence by demonstrating the prominent role of the Pacific Meridional Mode (PMM) on the inter-annual variability of CESR. A strong positive phase of PMM is linked to a significant rainfall deficit over CE (10--15%; p-value<0.05) during the boreal summer season. Further, we demonstrate the implications of this novel teleconnection pattern in improving the predictability of CESR at seasonal timescales -- along with the mutually exclusive and independent North Atlantic Oscillation (NAO). Our study thus reveals the far-reaching influences of the North Pacific Ocean on an improved prediction of the CESR variability, which may aid in improving the adaptation and mitigation strategies.

Just in the last four decades, ~5 million sq.km of humid land has transformed into a dryland. A rapidly warming climate is expected to further accelerate this dryland expansion. Consequently, not only will societies face permanent water insecurity, but plant and animal biodiversity will be under threat. Despite these consequences, very little is known about the physical mechanisms which cause irreversible drying of humid regions. As a result, our ability to predict the future expansion of drylands and its impact remains limited. Learning how drylands have expanded historically from humid regions could hold the key to predicting how they might expand in the future. So far, dryland expansion has been attributed to shifts in atmospheric circulation, topography, and orbital cycles. However, these changes occur at timescales reaching up to millions of years and thus do not fully explain the current rate of expansion. In this regard, the role of vegetation response to atmospheric drying and the consequent changes in land–atmosphere interactions have been largely ignored as possible mechanistic pathways of dryland expansion. Here, by tracking the air flowing over drylands, I show that the warming and drying of that air by changes in dryland vegetation-driven land–atmosphere feedback contributes to dryland expansion in the downwind direction. As they dry, drylands contribute less moisture and more heat to downwind humid regions, reducing precipitation and increasing atmospheric water demand, which ultimately causes their aridification. In ~40% of the land area that recently transitioned from a humid region into a dryland, self-expansion accounted for >50% of the observed aridification. Our results highlight the urgent need for climate change mitigation measures in drylands and provide a scientific basis for land-based interventions to prevent irreversible drying of humid regions

Water is indispensable for life on Earth. Plants use water either from recent precipitation (within a month) or from past precipitation stored in deeper soil (P_P; at least a month ago) to maintain metabolism and growth. It is widely known that plants tend to rely more on P_P to buffer against short-term rainfall deficits. However, how this reliance has changed in response to global change remains unclear. Here we develop a novel framework to estimate temporal changes in plant reliance on P_P during the past four decades. Observational data reveal that 42% of tropical and subtropical natural ecosystems have experienced a significant increase in plant reliance on P_P over the period 1982–2021 (P < 0.05). Such an increase is consistent with greening during the late growing season in drylands and drying during the wet-to-dry transitional period in non-drylands, when short-term precipitation fails to meet plant water demand. Adaptive changes in root depth and species composition may further facilitate this change in P_P reliance, especially in drylands. Our study highlights the importance of P_P in ecosystem functioning and implies an increasing ecosystem resilience to climate variability.

Rainfed crops account for approximately 40% of India’s food production and support 60% of its livestock. Although linked to oceanic monsoon rainfall, their productivity also depends on terrestrial evaporation, particularly in the non-monsoon season. However, the degree to which rainfed crops also rely on moisture sourced from upwind irrigated areas, remains largely unknown. Using a combination of models and observations, we show that upwind irrigated crops contribute 7%±6% of the rainfall over rainfed areas, rising to 15%±10% during the pre-monsoon months. In the absence of this input, water stress experienced by rainfed crops can increase by 5–10% during the crucial mid to late crop growth phases, potentially affecting yields. Our results reveal an unrecognized atmospheric link between irrigated and rainfed agriculture, which is overlooked in current agricultural policies. Planning and managing these systems in an holistic manner can help strengthen regional food and water security under future climates.

Understanding the partitioning of runoff into baseflow and quickflow is crucial for informed decision‐making in water resource management, guiding the implementation of flood mitigation strategies, and enhancing drought resilience measures. Methods that combine the physically based models with machine learning (ML) have demonstrated potential for global runoff estimation. However, such “hybrid” approaches remain unexplored for baseflow estimation. Here, we develop a ML approach combined by the physically‐based Budyko framework for baseflow estimation by incorporating the baseflow coefficient (BFC) curve as a physical constraint. Parameters of the original Budyko curve and the newly developed BFC curve are estimated based on 13 climatic and physiographic properties using boosted regression trees. BRT models are trained and tested in 1,461 catchments worldwide and subsequently applied to the entire global land surface at a 0.25° grid scale. The models exhibit strong performance during the testing phase, with R ² values of 0.96 and 0.91 for runoff and baseflow, respectively. Results indicate that, on average, 35.3% of global continental precipitation (819 mm yr ⁻¹ ) is partitioned into runoff (292 mm yr ⁻¹ ), comprising 19.6% as baseflow (162 mm yr ⁻¹ ) and 15.7% as quickflow (130 mm yr ⁻¹ ). Among the 13 climatic and physiographic properties analyzed, vegetation properties emerge as primary controls for Budyko parameter α while soil properties dominate parameter Q _{b , p} in BFC curve, although significant spatial variability is observed across global catchments. Overall, the proposed framework provides a global data set and methodology for runoff partitioning while revealing how catchment properties control this process.

Subseasonal prediction of heatwaves, although highly valuable for risk reduction, is challenging because heatwave onsets and propagation are complex processes with both fast and slow drivers from local to global scale. Traditionally, subseasonal forecasting relies heavily on dynamical model ensembles, which are complex and of high computational cost. As an alternative, machine learning provides potentially performant solutions that may match or even outperform these physical-based models. Transformers, in particular, are the current state-of-the-art deep learning infrastructures, and using multi-head-attention allows them to keep track of long-term complex dependencies in timeseries data. However, to better forecast heatwaves subseasonally, it is essential to move beyond purely predictor-to-target associative measures when identifying the sources of predictability. Such endeavours require causal frameworks that provide directionality and explainable power for the predictor-to-target relationships.This study seeks to implement the PCMCI+ (Runge, 2020) framework to identify causal drivers of heatwaves on the Iberian Peninsula on a subseasonal scale. Causally-selected predictors are employed to forecast the occurence of heatwaves up to six weeks in advance using transformer networks, both for different seasons and sub-regions in the Iberian Peninsula. Preliminary results reveal heatwaves can be predicted with reasonable accuracy with a forecast window of six weeks, particularly in water limited regions, using causality-based machine learning.Reference:Runge, J. (2020). Discovering contemporaneous and lagged causal relations in autocorrelated nonlinear time series datasets. In Conference on Uncertainty in Artificial Intelligence, pages 1388–1397. PMLR.