a – Improvement of agricultural and water management using satellite data
Early activities within DRY–2–DRY demonstrated that state-of-the-art satellite datasets of evaporation where insufficient to accomplish the first objective of providing evidence of the impacts of drought and heatwaves on terrestrial evaporation. That led to a large time investment to improve the GLEAM model (www.gleam.eu) aiming to better capture plant transpiration during conditions of stress. That required the application of the framework at increasingly higher resolutions by utilising data from the Sentinel constellation at 1 km resolution over continental scales, and at 100 m over regional scales. A longstanding collaboration with VanderSat was established to foster the development of GLEAM at hyper-resolution, in order to turn it into a very valuable asset for water management and agricultural irrigation planning. Spin-off activities included the collaboration with the European Space Agency (ESA) in several projects that feed into the EU Green Deal, the EU Data Strategy and the EU Destination Earth (DestinE), directed towards the modelling of a high-resolution digital replica of the Earth's hydrological cycle based on satellite observations.
b – Enabling a better representation of land processes in climate models
A central activity within DRY–2–DRY was to improve global evaporation and soil moisture data to benchmark climate model representation of land surface processes. The accurate model representation of evaporation is critical to predict droughts and heatwaves. Data produced during the project was used in the IPCC AR6 for the purpose of climate model benchmarking. The commitment to refining global satellite products for their use as benchmark in climate models led the DRY–2–DRY team to explore hybrid modelling, i.e. the combined use of process-based and machine learning algorithms. This demonstrated great potential to better understand land surface processes at global scales, including evaporation, soil moisture and plant water stress. Therefore, the project heralded an alternative way to study these variables and represent them in climate models. Likewise, earlier research within the project also demonstrated that novel satellite observations (e.g. of sun-induced fluorescence) can be used to estimate plant use of water. These lines of work leave a legacy beyond DRY–2–DRY. Updated datasets of these variables will continue being made available at www.gleam.eu which counted on more than 8,000 unique data users during the lifespan of DRY–2–DRY.
c – Uncovering whether the global water cycle is accelerating
A clear expectation derived from climate model simulations and basic thermodynamic principles is that precipitation will increase globally as the planet warms. The term 'water cycle acceleration' has been coined to refer to this process. This is not a homogenous response at regional scales, as it is conditioned on the ability of the land surface to evaporate more water as temperature rises. In fact, this is not always the case, and most dry regions are expected to become drier and receive less precipitation in the future. Yet globally, our research unequivocally shows that land evaporation will monotonically increase following global warming trends. Since 2015, our contributions to the annual State of the Climate Report from the Bulletin of the American Meteorological Society (BAMS) have confirmed these earlier findings. These recursive publications using data produced during DRY–2–DRY have consistently demonstrated that El Niño-driven droughts cause negative anomalies in global evaporation because they trigger droughts in the Southern Hemisphere, but decadal trends still follow global warming. This is an important finding that implies that the future water availability over the continents will be highly sensitive to greenhouse emissions, and the trends in the intensity of El Niño–La Niña cycles. The P.I.'s involvement within the evaporation scientific community led to a role as board member in the Global Climate Observing System (GCOS), where he is responsible for the 'Land Evaporation' Essential Climate Variable (ECV).
d – Better understanding the lethality of drought–heatwave compound events
Recent heatwaves frequently coincided with soil droughts, largely due to the feedbacks illustrated in Fig. 1. While this synchronisation with soil drought acts to intensify air temperature, it also lowers air humidity. Using the novel HAMSTER software tool developed during DRY–2–DRY, it was demonstrated that these compound dry–hot conditions are highly detrimental for ecosystem primary productivity as they lead to severe vegetation stress. However, for humans, the situation is different: the lowering of air humidity due to soil drought reduces heat stress, instead of increasing it as it occurs for plants. To understand the net impact of soil desiccation on the human morbidity and mortality of heatwaves, another software tool, called CLASS4GL, was developed within DRY–2–DRY. CLASS4GL combines weather balloon data, satellite observations and atmospheric modelling. The tool was used to disentangle the effects of land feedbacks on the local likelihood of rainfall during droughts and heatwaves. Then, combined with meta-analyses of morbidity and mortality, it was demonstrated that soil drought actually leads to a mild reduction in human heatwave lethality. This finding questioned the efficiency of traditionally proposed mitigation strategies involving surface moistening to increase evaporative cooling, and it gave rise to an interdisciplinary research line to explore the impact of drought on human physiology during heat strokes.
e – Predicting whether droughts will become more severe in the future
The DRY–2–DRY work also focused on future dry extremes, not just historical events. Work during the last stages of the project demonstrated that future drought duration is underestimated by the current generation of climate models. The relationship between the Coupled Model Intercomparison Project (CMIP5 and CMIP6) past estimates and future projections of meteorological drought duration was contrasted with past satellite observations, using what is often referred to as an 'emergent constraint'. Results showed that the 21st century global increase in dry-spell duration will be significantly larger than predicted by the CMIP5 and CMIP6 model ensembles. These findings revealed extensive world regions where global warming will cause meteorological drought aggravation, and they emphasised the importance of reducing climate model errors in precipitation, cloud formation and land–atmosphere feedbacks.
f – Revealing whether droughts and heatwaves can self-propagate
A large part of DRY–2–DRY was dedicated to unravelling the effects of land feedbacks on the expansion and concatenation of droughts and heatwaves. Results demonstrated that the paradigm proposed in DRY–2–DRY was indeed valid: droughts and heatwaves do self-propagate. As a soil drought strikes, evaporation is reduced, which contributes to the drying but also the warming of the air. Moreover, the atmosphere reacts in such a manner that it allows the accumulated heat and dry air to persist from day to day. As that warm and dry air flows downwind, it creates a propagation front, leading to the expansion and concatenation of drought conditions. This propagation is fuelled by the soil dryness caused by the drought itself; similar to the propagation of a wildfire. This process of drought 'self-propagation' is especially important in drylands, where it can account for ~20–30% of the precipitation deficits causing droughts – see Fig. 2 (Schumacher et al., 2022). We also demonstrated that this self-propagation applies to heatwaves as well: as heatwaves dry the surface, the enhanced warming over the land is transported downwind leading to their expansion and even to a concatenation of multiple heatwave events – see Fig. 3 (Schumacher et al. 2019). During recent European mega-heatwaves, over 30% of the heat anomaly was caused through this process.
g – Challenging the notion of 'watershed' as management unit in hydrology
Watersheds (i.e. hydrological catchments) are tightly interconnected via atmospheric water transfers as shown during DRY–2–DRY. This contradicts the traditional views of water resources and water scarcity as local aspects for which watershed‐based management appears adequate. The truth is that precipitation, as a main source of freshwater, depends on moisture supplied through land evaporation from outside the watershed, sometimes thousands of kilometres away. This notion of evaporation as a remote supply to local precipitation is typically not considered in hydrological water assessments. The work within DRY–2–DRY showed, for instance, that up to ¾ of summer precipitation over European watersheds depends on moisture supplied from other watersheds. In other words: watersheds are not autarkic systems. This emphasized the need for global water governance to secure freshwater availability. It also highlighted the role of upwind watershed evaporation for drought occurrence in downwind watersheds. Overall, this has far-reaching implications: the measures that are currently taken in the management of local land and water resources have quantifiable repercussions on water availability in remote regions, which requires consensual decisions, not only between different basins, but also between countries.
