Since 2010 an uninterrupted sequence of dry years, with annual rainfall deficits ranging from 25 to 45 %, has prevailed in central Chile (western South America, 30–38° S). Although intense 1- or 2-year droughts are recurrent in this Mediterranean-like region, the ongoing event stands out because of its longevity and large extent. The extraordinary character of the so-called central Chile megadrought (MD) was established against century long historical records and a millennial tree-ring reconstruction of regional precipitation. The largest MD-averaged rainfall relative anomalies occurred in the northern, semi-arid sector of central Chile, but the event was unprecedented to the south of 35° S. ENSO-neutral conditions have prevailed since 2011 (except for the strong El Niño in 2015), contrasting with La Niña conditions that often accompanied past droughts. The precipitation deficit diminished the Andean snowpack and resulted in amplified declines (up to 90 %) of river flow, reservoir volumes and groundwater levels along central Chile and westernmost Argentina. In some semi-arid basins we found a decrease in the runoff-to-rainfall coefficient. A substantial decrease in vegetation productivity occurred in the shrubland-dominated, northern sector, but a mix of greening and browning patches occurred farther south, where irrigated croplands and exotic forest plantations dominate. The ongoing warming in central Chile, making the MD one of the warmest 6-year periods on record, may have also contributed to such complex vegetation changes by increasing potential evapotranspiration. We also report some of the measures taken by the central government to relieve the MD effects and the public perception of this event. The understanding of the nature and biophysical impacts of the MD helps as a foundation for preparedness efforts to confront a dry, warm future regional climate scenario.

Risk management has reduced vulnerability to floods and droughts globally^1,2, yet their impacts are still increasing³. An improved understanding of the causes of changing impacts is therefore needed, but has been hampered by a lack of empirical data^4,5. On the basis of a global dataset of 45 pairs of events that occurred within the same area, we show that risk management generally reduces the impacts of floods and droughts but faces difficulties in reducing the impacts of unprecedented events of a magnitude not previously experienced. If the second event was much more hazardous than the first, its impact was almost always higher. This is because management was not designed to deal with such extreme events: for example, they exceeded the design levels of levees and reservoirs. In two success stories, the impact of the second, more hazardous, event was lower, as a result of improved risk management governance and high investment in integrated management. The observed difficulty of managing unprecedented events is alarming, given that more extreme hydrological events are projected owing to climate change³.

We introduce the first catchment dataset for large sample studies in Chile. This dataset includes 516 catchments; it covers particularly wide latitude (17.8 to 55.0° S) and elevation (0 to 6993 m a.s.l.) ranges, and it relies on multiple data sources (including ground data, remote-sensed products and reanalyses) to characterise the hydroclimatic conditions and landscape of a region where in situ measurements are scarce. For each catchment, the dataset provides boundaries, daily streamflow records and basin-averaged daily time series of precipitation (from one national and three global datasets), maximum, minimum and mean temperatures, potential evapotranspiration (PET; from two datasets), and snow water equivalent. We calculated hydro-climatological indices using these time series, and leveraged diverse data sources to extract topographic, geological and land cover features. Relying on publicly available reservoirs and water rights data for the country, we estimated the degree of anthropic intervention within the catchments. To facilitate the use of this dataset and promote common standards in large sample studies, we computed most catchment attributes introduced by Addor et al. (2017) in their Catchment Attributes and MEteorology for Large-sample Studies (CAMELS) dataset, and added several others. We used the dataset presented here (named CAMELS-CL) to characterise regional variations in hydroclimatic conditions over Chile and to explore how basin behaviour is influenced by catchment attributes and water extractions. Further, CAMELS-CL enabled us to analyse biases and uncertainties in basin-wide precipitation and PET. The characterisation of catchment water balances revealed large discrepancies between precipitation products in arid regions and a systematic precipitation underestimation in headwater mountain catchments (high elevations and steep slopes) over humid regions. We evaluated PET products based on ground data and found a fairly good performance of both products in humid regions (r > 0.91) and lower correlation (r < 0.76) in hyper-arid regions. Further, the satellite-based PET showed a consistent overestimation of observation-based PET. Finally, we explored local anomalies in catchment response by analysing the relationship between hydrological signatures and an attribute characterising the level of anthropic interventions. We showed that larger anthropic interventions are correlated with lower than normal annual flows, runoff ratios, elasticity of runoff with respect to precipitation, and flashiness of runoff, especially in arid catchments. CAMELS-CL provides unprecedented information on catchments in a region largely underrepresented in large sample studies. This effort is part of an international initiative to create multi-national large sample datasets freely available for the community. CAMELS-CL can be visualised from http://camels.cr2.cl and downloaded from https://doi.pangaea.de/10.1594/PANGAEA.894885.

The socio-ecological sensitivity to water deficits makes Chile highly vulnerable to global change. New evidence of a multi-decadal drying trend and the impacts of a persistent drought that since 2010 has affected several regions of the country, reinforce the need for clear diagnoses of the hydro-climate changes in Chile. Based on the analysis of long-term records (50+ years) of precipitation and streamflow, we confirm a tendency toward a dryer condition in central-southern Chile (30–48°S). We describe the geographical and seasonal character of this trend, as well as the associated large-scale circulation patterns. When a large ensemble of climate model simulations is contrasted to observations, anthropogenic forcing appears as the leading factor of precipitation change. In addition to a drying trend driven by greenhouse gas forcing in all seasons, our results indicate that the Antarctic stratospheric ozone depletion has played a major role in the summer rainfall decline. Although average model results agree well with the drying trend’s seasonal character, the observed change magnitude is two to three times larger than that simulated, indicating a potential underestimation of future projections for this region. Under present-day carbon emission rates, the drying pathway in Chile will likely prevail during the next decades, although the summer signal should weaken as a result of the gradual ozone layer recovery. The trends and scenarios shown here pose substantial stress on Chilean society and its institutions, and call for urgent action regarding adaptation measures.

A decade-long (2010–2020) period with precipitation deficits in central–south Chile (30–41^∘ S), the so-called megadrought (MD), has led to streamflow depletions of larger amplitude than expected from precipitation anomalies, indicating an intensification in drought propagation. We analysed the catchment characteristics and runoff mechanisms modulating such intensification by using the CAMELS-CL dataset and simulations from the HBV hydrological model. We compared annual precipitation–runoff (P–R) relationships before and during the MD across 106 basins with varying snow-/rainfall regimes and identified those catchments where drought propagation was intensified. Our results show that catchments' hydrological memory – modulated by snow and groundwater – is a key control of drought propagation. Snow-dominated catchments (30–35^∘ S) feature larger groundwater contribution to streamflow than pluvial basins, which we relate to the infiltration of snowmelt over the Western Andean Front. This leads to longer memory in these basins, represented by a significative correlation between autumn streamflow (when snow has already melted) and the precipitation from the preceding year. Hence, under persistent drought conditions, snow-dominated catchments accumulate the effects of precipitation deficits and progressively generate less water, compared with their historical behaviour, notably affecting central Chile, a region with limited water supply and which concentrates most of the country's population and water demands. Finally, we addressed a general question: what is worse – an extreme single-year drought or a persistent moderate drought? In snow-dominated basins, where water provision strongly depends on both the current and previous precipitation seasons, an extreme drought induces larger absolute streamflow deficits; however persistent deficits induce a more intensified propagation of the meteorological drought. Hence, the worst scenario would be an extreme meteorological drought following consecutive years of precipitation below average, as occurred in 2019. In pluvial basins of southern Chile (35–41^∘ S), hydrologic memory is still an important factor, but water supply is more strongly dependant on the meteorological conditions of the current year, and therefore an extreme drought would have a higher impact on water supply than a persistent but moderate drought.