Search

International audience; Global warming is expected to increase droughts and heatwaves, and consequently fire danger in southern Europe in the forthcoming decades. However, an assessment of the uncertainties associated with this general trend at regional scales, relevant to decision-making, is still missing. This study aims at assessing potential climate change impacts on fire danger over France through the projection of the widely used Fire Weather Index (FWI) and at quantifying the different sources of climate-driven uncertainty associated with these projections. We used daily climate experiments covering the 1995-2098 period under two scenarios (RCP4.5 and RCP8.5) provided by the EURO-CORDEX initiative. Our results show an overall increase in FWI throughout the century, with the largest absolute increases in the Mediterranean area. Model uncertainty was very high in western France, previously identified as a potential fire-prone region under future climate. In contrast, large increases in FWI in the Mediterranean area showed low uncertainty across models. Besides, analyzing the natural variability of FWI revealed that extreme years under present-day climate could become much more frequent by the end of the century. The FWI is projected to emerge from the background of natural variability by mid-twenty-first century with a summer elevated fire danger three times more likely when summer temperature anomaly exceeds + 2 degrees C.

Data are available in PROMÉTHÉE2: Forest fires database for Mediterranean area in France, and can be extracted at https://www.promethee.com/en/incendies using the following selection criteria: fire type, Forêt; date range, 1 January 1995 to 31 December 2018.; International audience; Modeling wildfire activity is crucial for informing science-based risk management and understanding the spatiotemporal dynamics of fire-prone ecosystems worldwide. Models help disentangle the relative influences of different factors, understand wildfire predictability, and provide insights into specific events. Here, we develop Firelihood, a two-component, Bayesian, hierarchically structured, probabilistic model of daily fire activity, which is modeled as the outcome of a marked point process: individual fires are the points (occurrence component), and fire sizes are the marks (size component). The space-time Poisson model for occurrence is adjusted to gridded fire counts using the integrated nested Laplace approximation (INLA) combined with the stochastic partial differential equation (SPDE) approach. The size model is based on piecewise-estimated Pareto and generalized Pareto distributions, adjusted with INLA. The Fire Weather Index (FWI) and forest area are the main explanatory variables. Temporal and spatial residuals are included to improve the consistency of the relationship between weather and fire occurrence. The posterior distribution of the Bayesian model provided 1,000 replications of fire activity that were compared with observations at various temporal and spatial scales in Mediterranean France. The number of fires larger than 1 ha across the region was coarsely reproduced at the daily scale, and was more accurately predicted on a weekly basis or longer. The regional weekly total number of larger fires (10-100 ha) was predicted as well, but the accuracy degraded with size, as the model uncertainty increased with event rareness. Local predictions of fire numbers or burned areas also required a longer aggregation period to maintain model accuracy. The estimation of fires larger than 1 ha was also consistent with observations during the extreme fire season of the 2003 unprecedented heat wave, but the model systematically underrepresented large fires and burned areas, which suggests that the FWI does not consistently rate the actual danger of large fire occurrence during heat waves. Firelihood enabled a novel analysis of the stochasticity underlying fire hazard, and offers a variety of applications, including fire hazard predictions for management and projections in the context of climate change.

It is widely recognized that future rainfall extremes will intensify. This expectation is tied to the Clausius-Clapeyron (CC) relation, stating that the maximum water vapour content in the atmosphere increases by 6–7% per degree warming. Scaling rates for the dependency of hourly precipitation extremes on near-surface (dew point) temperature derived from day-to-day variability have been found to exceed this relation (super-CC). However, both the applicability of this approach in a long-term climate change context, and the physical realism of super-CC rates have been questioned. Here, we analyse three different climate change experiments with a convection-permitting model over Western Europe: simple uniform-warming, 11-year pseudo-global warming and 11-year global climate model driven. The uniform-warming experiment results in consistent increases to the intensity of hourly rainfall extremes of approximately 11% per degree for moderate to high extremes. The other two, more realistic, experiments show smaller increases—usually at or below the CC rate—for moderate extremes, mostly resulting from significant decreases to rainfall occurrence. However, changes to the most extreme events are broadly consistent with 1.5–2 times the CC rate (10–14% per degree), as predicted from the present-day scaling rate for the highest percentiles. This result has important implications for climate adaptation. This article is part of a discussion meeting issue ‘Intensification of short-duration rainfall extremes and implications for flash flood risks’.

A large number of recent studies have aimed at understanding short-duration rainfall extremes, due to their impacts on flash floods, landslides and debris flows and potential for these to worsen with global warming. This has been led in a concerted international effort by the INTENSE Crosscutting Project of the GEWEX (Global Energy and Water Exchanges) Hydroclimatology Panel. Here, we summarize the main findings so far and suggest future directions for research, including: the benefits of convection-permitting climate modelling; towards understanding mechanisms of change; the usefulness of temperature-scaling relations; towards detecting and attributing extreme rainfall change; and the need for international coordination and collaboration. Evidence suggests that the intensity of long-duration (1 day+) heavy precipitation increases with climate warming close to the Clausius–Clapeyron (CC) rate (6–7% K −1 ), although large-scale circulation changes affect this response regionally. However, rare events can scale at higher rates, and localized heavy short-duration (hourly and sub-hourly) intensities can respond more strongly (e.g. 2 × CC instead of CC). Day-to-day scaling of short-duration intensities supports a higher scaling, with mechanisms proposed for this related to local-scale dynamics of convective storms, but its relevance to climate change is not clear. Uncertainty in changes to precipitation extremes remains and is influenced by many factors, including large-scale circulation, convective storm dynamics andstratification. Despite this, recent research has increased confidence in both the detectability and understanding of changes in various aspects of intense short-duration rainfall. To make further progress, the international coordination of datasets, model experiments and evaluations will be required, with consistent and standardized comparison methods and metrics, and recommendations are made for these frameworks. This article is part of a discussion meeting issue ‘Intensification of short-duration rainfall extremes and implications for flash flood risks’.

Short-duration (1–3 h) rainfall extremes can cause serious damage to societies through rapidly developing (flash) flooding and are determined by complex, multifaceted processes that are altering as Earth’s climate warms. In this Review, we examine evidence from observational, theoretical and modelling studies for the intensification of these rainfall extremes, the drivers and the impact on flash flooding. Both short-duration and long-duration (>1 day) rainfall extremes are intensifying with warming at a rate consistent with the increase in atmospheric moisture (~7% K−1), while in some regions, increases in short-duration extreme rainfall intensities are stronger than expected from moisture increases alone. These stronger local increases are related to feedbacks in convective clouds, but their exact role is uncertain because of the very small scales involved. Future extreme rainfall intensification is also modulated by changes to temperature stratification and large-scale atmospheric circulation. The latter remains a major source of uncertainty. Intensification of short-duration extremes has likely increased the incidence of flash flooding at local scales, and this can further compound with an increase in storm spatial footprint to considerably increase total event rainfall. These findings call for urgent climate change adaptation measures to manage increasing flood risks. Short-duration rainfall extremes are determined by complex processes that are affected by the warming climate. This Review assesses the evidence for the intensification of short-duration rainfall extremes, the associated drivers and the implications for flood risks.

[Departement_IRSTEA]Territoires [ADD1_IRSTEA]Adaptation des territoires au changement global; International audience; Large wildfires across parts of France can cause devastating damage which puts lives, infrastructure, and the natural ecosystem at risk. In the climate change context, it is essential to better understand how these large wildfires relate to weather and climate and how they might change in a warmer world. Such projections rely on the development of a robust modelling framework linking large wildfires to present-day atmospheric variability. Drawing from a MODIS product and a gridded meteorological dataset, we derived a suite of biophysical and fire danger indices and developed generalized linear models simulating the probability of large wildfires ( > 100 ha) at 8 km spatial and daily temporal resolutions across the entire country over the last two decades. The models were able to reproduce large-wildfire activity across a range of spatial and temporal scales. Different sensitivities to weather and climate were detected across different environmental regions. Long-term drought was found to be a significant predictor of large wildfires in flammability-limited systems such as the Alpine and south-western regions. In the Mediterranean, large wildfires were found to be associated with both short-term fire weather conditions and longer-term soil moisture deficits, collectively facilitating the occurrence of large wildfires. Simulated probabilities on days with large wildfires were on average 2-3 times higher than normal with respect to the mean seasonal cycle, highlighting the key role of atmospheric variability in wildfire spread. The model has wide applications, including improving our understanding of the drivers of large wildfires over the historical period and providing a basis on which to estimate future changes to large wildfires from climate scenarios.

[Departement_IRSTEA]Territoires [ADD1_IRSTEA]Adaptation des territoires au changement global; International audience; In recent years, numerous studies have investigated the relationship of extreme precipitation intensities with temperature derived from a binning method using present-day weather data. The aim of this approach is that these dependencies, or so-called apparent scaling rates, provide insights on how precipitation extremes could evolve in the future climate. In a recent paper, Bao et al. showed that there is a large discrepancy between the present-day apparent scaling rates

Climatological features of observed annual maximum hourly precipitation have not been documented systematically compared to those on daily timescales due to observational limitations. Drawing from a quality-controlled database of hourly records sampling different climatic regions including the United States, Australia, the British Isles, Japan, India and peninsular Malaysia over the 1950–2016 period, we examined climatological features of annual maximum precipitation (AMP) across timescales ranging from 1-hr (AMP1−hr) to 24-hr (AMP24−hr). Our analysis reveals strong relations between the magnitude of AMP and the climatological average annual precipitation (AAP), with geographic variations in the magnitude of AMP24−hr across topographic gradients not evident in AMP1−hr. Most AMP1−hr are found to be embedded within short-duration storms (>70% of AMP1−hr are embedded within 1–5 h storms), especially in regions with low AAP and in the tropical zone. Likewise, most AMP24−hr are found to be the accumulation of a very limited number of wet hours in the 24-h period (>80% of AMP24−hr are due to storms lasting

Although it has been documented that daily precipitation extremes are increasing worldwide, faster increases may be expected for subdaily extremes. Here after a careful quality control procedure, we compared trends in hourly and daily precipitation extremes using a large network of stations across the United States (U.S.) within the 1950–2011 period. A greater number of significant increasing trends in annual and seasonal maximum precipitation were detected from daily extremes, with the primary exception of wintertime. Our results also show that the mean percentage change in annual maximum daily precipitation across the U.S. per global warming degree is ~6.9% °C−1 (in agreement with the Clausius-Clapeyron rate) while lower sensitivities were observed for hourly extremes, suggesting that changes in the magnitude of subdaily extremes in response to global warming emerge more slowly than those for daily extremes in the climate record.

We examine wet events (WEs) defined from an hourly rainfall dataset based on 64 gauged observations across India (1969–2016). More than 90% of the WEs (accounting for nearly 60% of total rainfall) are found to last less than or equal to 5 h. WEs are then clustered into six canonical local-scale storm profiles (CanWE). The most frequent canonical type (CanWE#1 and #2) are associated with very short and nominal rainfall. The remaining canonical WEs can be grouped into two broad families: (i) CanWE#3 and #5 with short (usually less than or equal to 3–4 h), but very intense rainfall strongly phase-locked onto the diurnal cycle (initiation peaks in mid-afternoon) and probably related to isolated thunderstorms or small mesoscale convective clusters (MCS), and (ii) CanWE#4 and #6 with longer and lighter rainfall in mean (but not necessarily for their maximum) and more independent of the diurnal cycle, thus probably related to larger MCSs or tropical lows. The spatial extent of the total rainfall received during each CanWE, as shown by IMERG gridded rainfall, is indeed smaller for CanWE#3 and #5 than for CanWE#4 and especially #6. Most of the annual maximum 1 hour rainfalls occur during CanWE#5. Long-term trend analysis of the June–September canonical WEs across boreal monsoonal India reveals an increase in the relative frequency of the convective storm types CanWE#3 and #5 in recent years, as expected from global warming and thermodynamic considerations.This article is part of a discussion meeting issue ‘Intensification of short-duration rainfall extremes and implications for flash flood risks’.

International audience; Anthropogenic climate change is leading to the intensification of extreme rainfall due to an increase in atmospheric water holding capacity at higher temperatures as governed by the Clausius-Clapeyron (C-C) relationship. However, the rainfall-temperature sensitivity (termed scaling) often deviates from the CC relationship. This manuscript uses classifications prescribed by regional-scale atmospheric circulation patterns to investigate whether deviations from the CC relationship in tropical Australia can be explained by differing weather types (WT). We show that the rainfall-temperature scaling differs depending on the WTs, with the difference increasing with rainfall magnitude. All monsoonal WTs have similar scaling, in excess of the CC relationship, while trade winds (the driest WTs) result in the greatest scaling, up to twice that of the CC relationship. Finally, we show the scaling for each WT also varies spatially, illustrating that both local factors and the WT will contribute to the behaviour of rainfall under warming.

[Departement_IRSTEA]Territoires [ADD1_IRSTEA]Adaptation des territoires au changement global; International audience; Changes in global fire activity are influenced by a multitude of factors including land-cover change, policies, and climatic conditions. This study uses 17 climate models to evaluate when changes in fire weather, as realized through the Fire Weather Index, emerge from the expected range of internal variability due to anthropogenic climate change using the time of emergence framework. Anthropogenic increases in extreme Fire Weather Index days emerge for 22% of burnable land area globally by 2019, including much of the Mediterranean and the Amazon. By the midtwenty-first century, emergence among the different Fire Weather Index metrics occurs for 33-62% of burnable lands. Emergence of heightened fire weather becomes more widespread as a function of global temperature change. At 2 °C above pre-industrial levels, the area of emergence is half that for 3 °C. These results highlight increases in fire weather conditions with human-caused climate change and incentivize local adaptation efforts to limit detrimental fire impacts.

[Departement_IRSTEA]Territoires [ADD1_IRSTEA]Dynamique et fonctionnement des écosystèmes [Departement_IRSTEA]Territoires [ADD1_IRSTEA]Dynamique et fonctionnement des écosystèmes; International audience; Six weather types (WTs) are computed for tropical Australia during the wet season (November-March 1979-2015) using cluster analysis of 6-hourly low-level winds at 850 hPa. The WTs may be interpreted as a varying combination of at least five distinct phenomena operating at different time scales: the diurnal cycle, fast and recurrent atmospheric phenomena such as transient low pressure, the intraseasonal Madden-Julian oscillation, the annual cycle, and interannual variations mostly associated with El Niño-Southern Oscillation. The WTs are also strongly phase-locked onto the break/active phases of the monsoon; two WTs characterize mostly the trade-wind regime prevalent either at the start and the end of the monsoon or during its breaks, while three monsoonal WTs occur mostly during its core and active phases. The WT influence is strongest for the frequency of wet spells, while the influence on intensity varies according to the temporal aggregation of the rainfall. At hourly time scale, the climatological mean wet intensity tends to be near-constant in space and not systematically larger for the monsoonal WTs compared to other WTs. Nevertheless, one transitional WT, most prevalent around late November and characterized by weak synoptic forcings and overall drier conditions than the monsoonal WTs, is associated with an increased number of high hourly rainfall intensities for some stations, including for the interior of the Cape York Peninsula. When the temporal aggregation exceeds 6-12 h, the mean intensity tends to be larger for some of the monsoonal WTs, in association with more frequent and also slightly longer wet spells.

[Departement_IRSTEA]Territoires [ADD1_IRSTEA]Adaptation des territoires au changement global; International audience; Midlatitude synoptic weather regimes account for a substantial portion of annual precipitation accumulation as well as multi-day precipitation extremes across parts of the United States (US). However, little attention has been devoted to understanding how synoptic-scale patterns contribute to hourly precipitation extremes. A majority of 1-h annual maximum precipitation (AMP) across the western US were found to be linked to two coherent midlatitude synoptic patterns: disturbances propagating along the jet stream, and cutoff upper-level lows. The influence of these two patterns on 1-h AMP varies geographically. Over 95% of 1-h AMP along the western coastal US were coincident with progressive midlatitude waves embedded within the jet stream, while over 30% of 1-h AMP across the interior western US were coincident with cutoff lows. Between 30-60% of 1-h AMP were coincident with the jet stream across the Ohio River Valley and southeastern US, whereas a a majority of 1-h AMP over the rest of central and eastern US were not found to be associated with either midlatitude synoptic features. Composite analyses for 1-h AMP days coincident to cutoff lows and jet stream show that an anomalous moisture flux and upper-level dynamics are responsible for initiating instability and setting up an environment conducive to 1-h AMP events. While hourly precipitation extremes are generally thought to be purely convective in nature, this study shows that large-scale dynamics and baroclinic disturbances may also contribute to precipitation extremes on sub-daily timescales.

Temperature scaling studies suggest that hourly rainfall magnitudes might increase beyond thermodynamic expectations with global warming1–3; that is, above the Clausius–Clapeyron (CC) rate of ~6.5% °C−1. However, there is limited evidence of such increases in long-term observations. Here, we calculate continental-average changes in the magnitude and frequency of extreme hourly and daily rainfall observations from Australia over the years 1990–2013 and 1966–1989. Observed changes are compared with the uncertainty from natural variability and expected changes from CC scaling as a result of global mean surface temperature change. We show that increases in daily rainfall extremes are consistent with CC scaling, but are within the range of natural variability. In contrast, changes in the magnitude of hourly rainfall extremes are close to or exceed double the expected CC scaling, and are above the range of natural variability, exceeding CC × 3 in the tropical region (north of 23° S). These continental-scale changes in extreme rainfall are not explained by changes in the El Nino–Southern Oscillation or changes in the seasonality of extremes. Our results indicate that CC scaling on temperature provides a severe underestimate of observed changes in hourly rainfall extremes in Australia, with implications for assessing the impacts of extreme rainfall. Theory predicts that hourly rainfall extremes may increase with anthropogenic warming. Observations from Australia suggest changes two to three times above the Clausius–Clapeyron rate, above that expected from natural variability.

[Departement_IRSTEA]Territoires [ADD1_IRSTEA]Adaptation des territoires au changement global; National audience; Historical in situ sub-daily rainfall observations are essential for the understanding of short-duration rainfall extremes but records are typically not readily accessible and data are often subject to errors and inhomogeneities. Furthermore, these events are poorly quantified in projections of future climate change making adaptation to the risk of flash flooding problematic. Consequently, knowledge of the processes contributing to intense, short-duration rainfall is less complete compared with those on daily timescales. The INTENSE project is addressing this global challenge by undertaking a data collection initiative that is coupled with advances in high-resolution climate modelling to better understand key processes and likely future change. The project has so far acquired data from over 23 000 rain gauges for its global sub-daily rainfall dataset (GSDR) and has provided evidence of an intensification of hourly extremes over the US. Studies of these observations, combined with model simulations, will continue to advance our understanding of the role of local-scale thermodynamics and large-scale atmospheric circulation in the generation of these events and how these might change in the future.

An empirical statistical scheme for predicting September–December fires in New Caledonia in the southwestern Pacific Ocean region using a cross-validated generalized linear model has been developed for the 2000–10 period. The predictor employs July sea surface temperatures (SST) recorded over the Niño-4 box (5°S–5°N, 160°–210°E), which are closely related to austral spring (September–November) rainfall anomalies across New Caledonia. The correlation between the logarithm of observed and simulated total burned areas across New Caledonia is 0.87. A decrease in the local-scale skill (median correlation between the log of observed and simulated total burned areas in a 20-km radius around a rain gauge = 0.46) around the main town (Nouméa) and its suburbs in the southwest of Grande Terre, and also in northern New Caledonia, could be associated either with a weaker climatic forcing from the Niño-4 SST index or a small-scale climatic forcing not linearly related to the El Niño–Southern Oscillation (ENSO) phenomenon. It is more likely that the decrease is tied to the influence of human-driven factors that blur the regional-scale climatic signal mostly associated with central Pacific ENSO events.

International audience; Daily rainfall occurrence and amount at 55 stations over New Caledonia (NC, 20A degrees S, 166A degrees E) are examined throughout the calendar year during 1980-2010 using a Hidden Markov Model (HMM). Daily rainfall variability is described in terms of six discrete rainfall states identified by the HMM. Three states are interpreted as trade wind regimes associated with persistent low level anticyclonic conditions and weak to strong easterlies. The most prevalent state (state 1; 36 % of days) is dry everywhere and is characterized by an elongated anticyclone centered around Australia; NC is located on the eastern edge of this anticyclone. This state is predominant from mid-May to mid-December and peaks in September. The second most prevalent state (state 2; 27 % of days) and the last trade regime (state 4; 12 % of days) are most frequent in austral summer. States 2 and 4 are associated with the subtropical anticyclone centered south of NC, close to its climatological location in austral summer, and light (state 2) to moderate (state 4) rainfall mostly along the windward coast. A distinct state (state 3; 11 % of days) is mostly associated with wintertime extratropical eastward traveling troughs between New Caledonia and New Zealand, inducing light rainfall over the SW of the main island of New Caledonia. The two last states 5 and 6 are infrequent (< 15 % of days) but account for almost 70 % of total rainfall across the archipelago. They occur mostly in austral summer with their frequency of occurrence peaking in late February-early March. These states are associated with synoptic wave traveling eastward from eastern Australia leading to a strong influx of moisture from the equatorial latitudes when the associated cyclonic centre is located west of New Caledonia. The wettest state is also strongly modulated by intra-seasonal 15-80 days variability. These events contribute to intermittent southwestward shifts of the South Pacific Convergence Zone from its mean location northeast of New Caledonia. On interannual timescales, the occurrence of the rainfall states is modulated by El Nio Southern Oscillation (ENSO) events. The ENSO impact is strongest from mid-August to March with more (less) frequent trade regimes 1-2 and less (more) frequent unstable trade regime 4 and states 5-6 during central-Pacific warm (cold) ENSO events. Stochastic simulations of daily rainfall occurrence and amount at the 55 stations are generated by using predictors based on Nio 4 sea surface temperature index and a local intra-seasonal (15-80 days) OLR index superimposed on a climatological annual cycle. The cross-validated skill peaks in September-November.

[1] New Caledonia (NC; ∼166°E, 22°S) rainfall anomalies are more sensitive to central Pacific (CP) El Nino and La Nina events than to those exhibiting highest sea surface temperature (SST) anomalies in the eastern Pacific (EP). The linear relationship between NC rainfall anomalies and CP SST indices peaks from September to March (S–M). The seasonal S–M atmospheric anomalies observed in the South West (SW) Pacific during the warm CP events are highly dissimilar to the EP ones, while there are more similarities during the cold events with a higher amplitude during the CP ones. The warm CP events strengthen the southern Hadley cell around NC longitudes, with positive rainfall anomalies in the equatorial Pacific leading to an anomalous release of latent heat in the upper troposphere and an increased subsidence in the SW Pacific. Atmospheric anomalies are strongest in September–November because of a combination of a rather strong zonal SST gradient with the warmest SST in the equatorial Pacific just west of the dateline. The cold CP and EP events are associated with a southwestward shift of the South Pacific Convergence Zone with strongest atmospheric anomalies during the CP events. Squared wavelet coherence between NC rainfall and Nino 4 SST index shows that their negative correlations are mostly carried by two distinct timescales: the classical El Nino–Southern Oscillation (i.e., 3–6 years) variability and a quasi-decadal one (i.e., 10–12 years). The high-frequency (>1/8 cycle per year) correlations peak around Christmas and are quasi-stationary since 1950, whereas the low-frequency ones (

[1] We examined the relationships between fires detected by the Along Track Scanning Radiometer, algorithm 2 (ATSR-2) and the Moderate Resolution Imaging Spectrometer (MODIS) sensors and local-scale atmospheric conditions in New Caledonia (SW Pacific ∼165.5°E, 21.5°S) during the 1996–2008 (ATSR) and 2000–2008 (MODIS) time periods. A total of 3707 (MODIS) hotspots, representing 949 distinct fires, and 83 (ATSR) hotspots were observed during the study period. The annual mean frequency of hotspots peaks around the transition between the dry-cool season and the wet-warm season, that is, in September–December, when dry soils and fuels could be combined with increasing temperatures. The antecedent local-scale maximum temperature and rainfall anomalies recorded at the closest meteorological station were analyzed. While the signal in maximum temperature is weak and not robust among the fire records, the local-scale anomalies of rainfall are always clearly negative for at least 3 months before the fires (i.e., between June and December). The Effective Drought Index (EDI), based on rainfall only, and the Fire Weather Index (FWI), which combines rainfall, temperature, relative humidity, and wind speed, show similar behaviors before fires. This suggests that a simple rainfall index, as well as a more comprehensive one, is able to diagnose fire risks. At interannual time scale, cross-correlation analysis reveals that the seasonal June–August Nino 3.4 sea surface temperature index is strongly correlated (r = 0.78 for a second-order polynomial fit) with the monthly frequency of all MODIS hotspots in September–December, suggesting a strong potential predictability of fire variations in New Caledonia with a lead time of 1–4 months.