1. Introduction
Deep convective storms over Europe frequently produce heavy rainfall that poses a significant threat to life and property; Taszarek et al. (2019a) illustrated that heavy convective precipitation is the most common severe weather type over the continent. Regardless, forecasting this societally impactful convective storm hazard remains one of the most challenging (P. Groenemeijer, European Severe Storms Laboratory, 2022, personal communication). Convective storm development and evolution across the continent are often unanticipated as storm behavior is influenced by the regional topography (Güttler et al. 2015; Davolio et al. 2016; Katona et al. 2016) and surrounding seas (Azorin-Molina and Chen 2009; del Moral et al. 2020). Further, thunderstorms occur most frequently over several of these topographic features, including the Alps, Carpathians, mountainous Balkan Peninsula, and Mediterranean Sea (Taszarek et al. 2019a). In this study, we quantify the annual, seasonal, and diurnal spatiotemporal patterns of pan-European convective precipitation. Such awareness may lend itself to improved forecasts and preparedness for the associated flooding.
Past research on European convective storms has quantified the climatological synoptic-scale environments that support pan-European thunderstorm development (van Delden 2001; Mohr et al. 2019, 2020). For example, blocking high pressure systems over the Baltic Sea increases convective storm activity over western and central Europe, as unstable air is advected northward into the region by southwest flow on their western flanks (Mohr et al. 2019). For hazards, however, the bulk of the studies focused on the tornadoes, hail, and severe winds produced by these storms through analyses of the accompanying environments (Taszarek et al. 2020b) or individual case studies (Bechini et al. 2001; Bertato et al. 2003; Rebora et al. 2013; Holzer et al. 2018; Kunz et al. 2018; Pucillo et al. 2020; Taszarek et al. 2019b; Wilhelm et al. 2020).
While some prior studies have quantified precipitation patterns over the continent, many studies explored all heavy rain-producing atmospheric phenomena, rather than focusing on convective storms, and/or focused on convective precipitation over limited spatial domains (e.g., Frei and Schär 1998; Godart et al. 2011; Rebora et al. 2013; Fairman et al. 2017; Llasat et al. 2021). Rain gauges have been the primary data source for studies quantifying overland precipitation. Based on regional rain gauge data from 1970 to 1990, the annual mean precipitation over the Alps maximizes along the northern and southern rims with generally dry conditions over the adjacent flatland areas (Frei and Schär 1998). More recently, analysis of 5-km spatial resolution gridded rain gauge data identified the Piedmont–Ticino–Lombardy and Julian/Carnic Alps as the most active regions for heavy precipitation in the Alps annually (Isotta et al. 2014). Studies indicate that the Alps’ highest elevation peaks receive the greatest mean annual precipitation, and it is hypothesized that the orography and proximity to the sea support the generation of these maxima (Pavan et al. 2019). Godart et al. (2011) synthesized rain gauge and sounding data to evaluate heavy rain associated with banded precipitation features over the Cévennes–Vivarais region in France and found that stationary banded orographic rainfall events contribute up to 40% of the total annual precipitation (Godart et al. 2011). Llasat et al. (2021) specifically examined convective precipitation over two regions in the eastern Iberian Peninsula through gauge data; 16% of the total annual precipitation is from only 3%–6% of the convective events, with convective precipitation maximizing in late summer. As for the Alps, the authors hypothesized that the orography and sea are important features in generating these precipitation enhancements. While these studies are important contributions toward understanding the pan-European distribution of heavy convective precipitation, rain gauge data have limitations including inconsistencies of measurement devices and practices, and gridded gauge data have interpolation errors (Isotta et al. 2014).
Other studies leveraged satellite precipitation products to quantify precipitation patterns over marine regions such as the Mediterranean Sea. An analysis of the Tropical Rainfall Measuring Mission (TRMM) precipitation product over the period 1988–2007, at 0.25° horizontal and 3-hourly temporal resolution, revealed the western Mediterranean as the most active portion of the basin, with annual precipitation maxima during November and December (Mehta and Yang 2008). Complementarily, analysis of the Advanced Microwave Sounding Unit-B (AMSU-B) brightness temperatures, gridded to 0.2° horizontal resolution, highlighted a maximum in deep convection over the Mediterranean during autumn (Alhammoud et al. 2014). Based on AMSU-B, convection exhibited a weak diurnal cycle, with a slight preference for the early morning; however, AMSU-B is a polar-orbiting satellite, meaning it passes over a given location only twice daily, which may have biased the analysis.
Generally, prior research on European precipitation patterns has analyzed total precipitation as opposed to isolating convective precipitation or evaluated individual convective events. Often, rain gauge data are used for the analysis of localized overland regions or relatively spatiotemporally coarse satellite products for the analysis of oversea precipitation. This study is the first to leverage multiple datasets to quantify annual, seasonal, and diurnal heavy convective precipitation patterns over the European continent and the surrounding seas and oceans (Fig. 1).
Terrain map (m; shaded) including (a) labels for regions discussed in this study and (b) OPERA network radar locations, with the location of C-band radars (black dots), X-band radars (green dots), and S-band radars (purple dots) provided. Note shapefiles are unavailable for regions north of 60°N.
Citation: Monthly Weather Review 152, 7; 10.1175/MWR-D-23-0156.1
2. Data and methods
A combination of satellite, gridded surface station, ground-based radar, and reanalysis precipitation products, in conjunction with lightning data, is used to quantify annual and seasonal heavy convective precipitation spatial patterns and variations over the diurnal cycle. While each product has its own biases and limitations, utilizing multiple products allows us to cross-validate results and leverage the strengths of each, leading to a more robust climatological analysis (Taszarek et al. 2019a).
a. Precipitation
1) IMERG rain rate
The Integrated Multi-satellitE Retrievals for Global Precipitation Measurement (GPM) (IMERG) rain-rate (mm 30 min−1) precipitation product is obtained from the NASA GPM web portal (https://gpm.nasa.gov/data/imerg), for a 10-yr analysis period ranging from 2011 to 2020. IMERG products are available for the majority of the globe, including the European mountainous landscapes and surrounding seas. Thus, IMERG is the primary precipitation product used in this study. The rain-rate product is created using retrievals from the GPM satellite constellation after March 2014 (Hou et al. 2014) and the TRMM (Kummerow et al. 2000) satellite for months prior. This study uses the IMERG final run product (hereafter IMERG), the research-grade version of the algorithm, which incorporates Global Precipitation Climatology Centre (GPCC; Becker et al. 2013) precipitation gauge data, a climatological adjustment, and a quasi-Lagrangian time interpolation. Monthly GPCC precipitation gauge data are used to reduce regional-scale biases in the satellite precipitation estimates. These data are available at 30-min temporal and 0.1° horizontal spatial resolution, with “full” coverage available over a latitudinal range of 60°N–60°S. Only “partial” coverage is available at the highest latitudes (60°–90°N; 60°–90°S), and thus, the IMERG product is not used to analyze convective precipitation over the most northern regions of Europe. IMERG is forgiving of dropouts in individual sensors, and extended dropouts of sensors are rare (Huffman et al. 2020b). Additional details are provided in the IMERG technical documentation (Huffman et al. 2020a).
IMERG precipitation products have some limitations over mountainous and coastal regions, which should be considered when interpreting the precipitation signals. Blocky transitions in precipitation exist along coastlines with large land–sea gradients in precipitation magnitude (e.g., heavy precipitation over land transitioning to modest over sea; Navarro et al. 2019; Huffman et al. 2020a). Comparisons to surface station rain gauge data illustrate lower values over mountainous regions in IMERG products (Prakash et al. 2018; Ramsauer et al. 2018; Navarro et al. 2019; Tapiador et al. 2020). However, limitations in rain gauge products, including the spatial smoothing and resulting error in precipitation amounts (Fibbi et al. 2016; Navarro et al. 2019; Tramblay et al. 2019), should be considered. Further, caution should be used in comparisons between point-location rain gauge products and spatially coherent remote sensing products as they capture two different aspects of the precipitation field (Illingworth and Thompson 2022). Despite these limitations, IMERG offers continuous spatial and temporal precipitation information over the full study domain.
2) E-OBS surface mean precipitation
The European E-OBS daily surface station gridded climate dataset (Haylock et al. 2008; Cornes et al. 2018) for the 2011–20 period is an additional source of precipitation information used for comparison to IMERG. E-OBS v25.0e daily ensemble mean precipitation (mm) is available through the C3S Climate Data Store (CDS; Buontempo et al. 2020) at 0.1° spatial and daily temporal resolution. The station data, provided by various meteorological services across Europe and the Middle East, are quality controlled by the national agencies and quality checked by the European Climate Assessment and Dataset (Klok and Klein Tank 2009). E-OBS uncertainty is estimated using a stochastic 20-member ensemble, with the ensemble mean provided as the best estimate (e.g., Cornes et al. 2018; Frei and Isotta 2019). Thus, it is better practice to interpret E-OBS precipitation as an area mean over a given grid point rather than a point value (Bandhauer et al. 2022). Though beneficial, E-OBS information is confined to land surfaces only and lacks information over the European seas. Additionally, while E-OBS has the benefit of spatial and temporal continuity, the gridded product interpolates over large distances in some places and over complex topography with heterogeneous precipitation patterns, which may result in misleading accumulation values (Navarro et al. 2019).
3) OPERA rainfall rate
The Operational Programme for the Exchange of Weather Radar Information (OPERA; Huuskonen et al. 2014) is the radar program of the European Meteorological Network (EUMETNET), a network of 31 European National Meteorological and Hydrological Services (NMHS). OPERA has developed and offers composite pan-European radar-derived precipitation products, constructed from data gathered by a network of 211 primarily C-band ground-based radars located across the European continent (Fig. 1b). OPERA rainfall rate data are available at 15-min temporal and 2-km spatial resolution for the 2012–20 period. This study only utilizes data from 2014 to 2020 due to rounding issues in the OPERA BUFR files prior to 18 February 2013 (Lopez 2014).
There are several limitations of the OPERA ground-based precipitation product. Radar-derived precipitation rates may be contaminated with nonmeteorological targets such as insects, ground clutter, and radio interference, although clutter removal processing has improved this issue since its implementation in 2015 (Huuskonen et al. 2014; Saltikoff 2015). Ground-based radar products provide information over land and potentially nearshore coastal waters. They lack information over the sea due to the finite range of the radar beam and may have limited utility in mountainous regions due to beam blockage. OPERA products are comprised of data gathered from a heterogeneous radar network, consisting of 1) C-band, S-band, and X-band radars, 2) a 40-yr difference between the oldest and newest radars in the network resulting in a wide range of technologies across radars, and 3) instrumentation parts provided by a variety of manufacturers (Saltikoff et al. 2019). These heterogeneities pose challenges to the development of a unified product due to hurdles in managing radar hardware updates, as well as the wide range of scan strategies, data processing algorithms, and calibration among different countries (Peura et al. 2017; Saltikoff et al. 2017). Additionally, the precipitation rates are determined using a single Z–R relationship, assuming the Marshall and Palmer drop size distribution (Park et al. 2019). Consequently, OPERA products “underestimate” precipitation, as compared to the WMO surface synoptic observations (SYNOP; Park et al. 2019). Finally, coverage is constrained by the number of ground-based radars available across the region and the distance between them. More importantly, not all European countries participate in the OPERA program, resulting in an absence of data over portions of the continent.
4) EURADCLIM precipitation accumulation
The European climatological gauge-adjusted radar precipitation dataset (EURADCLIM; Overeem et al. 2023), available from the Royal Netherlands Meteorological Institute Ministry of Infrastructure and Water Management and accessible through the Koninklijk Nederlands Meteorologisch Instituut (KNMI) Data Platform, combines ground-based radar and surface station data. EURADCLIM is based on the EUMETNET OPERA gridded 15-min instantaneous surface rainfall rates. For EURADCLIM, a satellite-based cloud-type clutter filter and statistical filters remove additional nonmeteorological echoes from OPERA. Following, these data are merged with European Climate Assessment and Dataset (ECA&D) rain gauge data from ∼7700 stations from the NMHS using a Barnes objective analysis-based algorithm. EURADCLIM 1-h precipitation accumulation data are available on a 2-km horizontal spatial grid for the 8-yr period ranging 2013–20 (Overeem et al. 2022). EURADCLIM provides precipitation information over the European continent and some of its nearshore coastal waters.
5) ERA5 precipitation and convective precipitation
The fifth major global reanalysis produced by ECMWF (ERA5) total precipitation, convective precipitation, and convective rain-rate products are available at 0.25° spatial and hourly temporal resolution from 1940 to the present. Reanalyses, or datasets combining numerical modeling with data assimilation of quality-controlled observations, fill the gap of long-term observations over marine and other less-observed regions (Campos et al. 2022). ERA5 1-h total precipitation is the sum of the large-scale precipitation generated by the ECMWF Integrated Forecasting System (IFS) cloud parameterization scheme and the convective precipitation generated by the convective parameterization scheme, accumulated at the surface over a 1-h period. ERA5 1-h convective precipitation is the precipitation generated by the ECMWF IFS convective scheme only, accumulated at the surface over a 1-h period. ERA5 convective rain rate is the rainfall intensity as if it were spread evenly over the surface of a grid box, with 1 kg of water spread over 1 m2 equivalent to 1-mm depth. For our analysis, convective precipitation and convective rain rate were compared with minimal differences, and thus, only the convective precipitation product is presented in this study. These precipitation data provide another source of precipitation information, including convective precipitation, over the European seas. Care should be taken when comparing these values to some overland observations, as some observations provide information at a local point while model parameters are averages over a grid box.
b. Lightning
The Arrival Time Difference Network (ATDnet) instantaneous lightning flashes provided at 0.2° resolution and regridded to 0.1° are used to support the identification of convective precipitation for the period of interest. The latitudinal range of the dataset extends to 60°N, as for the IMERG “full coverage” product. Operated by and available from the Met Office, ATDnet is a long-range very-low-frequency (VLF) European lightning detection network, with the most recent version consisting of 11 sensors in a domain encompassing Europe, northern Africa, and part of the North Atlantic (Anderson and Klugmann 2014). The sensors detect sferics, electromagnetic signals with frequencies in the VLF range produced by cloud-to-ground lightning strikes and some strong intracloud pulses (Rakov and Uman 2003), with relatively continuous detection efficiency (Anderson and Klugmann 2014). ATDnet is capable of detecting 24% of intracloud flashes and 89% of cloud-to-ground flashes (Enno et al. 2016).
c. Identification of convective precipitation
Convective precipitation is objectively determined through a combination of IMERG and ATDnet datasets. A grid point is determined to have intense convective precipitation if 1) the IMERG precipitation rate is ≥10 mm h−1 and 2) an ATDnet lightning flash occurs within 20-km distance radially away from the precipitation grid point and within ±30 min of the precipitation time. A similar complementary analysis is performed using the EURADCLIM precipitation dataset, with the same precipitation threshold and lightning proximity requirements. The 10 mm h−1 precipitation rate is classified as convective based on known dBZ conversions (e.g., Battan and Theiss 1973). For convective rainfall, rainfall rate and Z are related through the expression Z = 300R1.4, where R is the rain rate (mm h−1). This relationship is used in the U.S. network of WSR-88D radars; note, however, that there are no objective criteria for choosing an optimal relation for a given time and location. The known expression to convert R to dBZ is given by dBZ = 10 log10Z. Thus, a 10 mm h−1 rain rate translates to ∼39-dBZ radar reflectivity. A radar reflectivity value exceeding 30–40 dBZ is a commonly used threshold to classify precipitation as convective in convective storm climatologies (e.g., Tokay and Short 1996; Parker and Knievel 2005; Gallus et al. 2008; Lombardo and Colle 2010). We acknowledge that 1) area distortion exists in the gridded precipitation products which may latitudinally bias their rain-rate values and 2) the horizontal grid spacing of the gridded precipitation products will impact their rain-rate values.
3. Annual precipitation patterns
a. Annual total precipitation accumulation patterns
Pan-European accumulated precipitation patterns are first evaluated through 10-yr annually averaged spatial maps of IMERG precipitation rate, E-OBS daily mean precipitation, EURADCLIM hourly precipitation accumulation, OPERA rainfall rate, and ERA5 hourly total precipitation (Fig. 2). IMERG accumulated precipitation exceeds 1500 mm yr−1 over many of the mountainous and coastal regions (Fig. 2a). In particular, accumulated precipitation is large over the Alps, the Dinaric Alps, and the coastal mountains of northwestern Spain and northern Portugal. Offshore, accumulated precipitation exceeds 1500 mm yr−1 over the coastal Adriatic Sea along the Balkan Peninsula and Italy, the coastal Ionian Sea, the Mediterranean Sea along the western Italian coastline, the southern North Sea, and the coastal Bay of Biscay along northern Spain. Additionally, IMERG accumulated precipitation is particularly large, >1750 mm yr−1, in the northwestern part of the analysis domain, including the northeastern Atlantic Ocean, as well as the western coasts of Great Britain, Ireland, and southern Norway.
Plan view of annually averaged precipitation accumulation (mm yr−1) for the 10-yr period ranging 2011–20 for (a) IMERG, (b) E-OBS, (e) ERA5 total precipitation, for the 8-yr period ranging 2013–20 for (c) EURADCLIM, and for the 7-yr period ranging 2014–20 for (d) OPERA.
Citation: Monthly Weather Review 152, 7; 10.1175/MWR-D-23-0156.1
Over land, E-OBS provides more detailed information of precipitation spatial patterns over the mountainous regions and indicates larger annual precipitation accumulation values over a number of localized areas within the mountains (Fig. 2b). Accumulation values exceed 2000 mm yr−1 over the Alpine peaks, the Apennines in the Riviera del Levante of Italy, the Julian Prealps, coastal Norway, and western Great Britain and Ireland. Precipitation enhancements over the inland mountains of Germany, Poland, Czechia, and Great Britain are also more apparent in E-OBS than in IMERG. Further, E-OBS illustrates that the large precipitation accumulation over coastal Norway observed in IMERG extends into the higher latitudes north of 60°N. Overall, however, E-OBS reveals similar spatial patterns in total precipitation and locations of relative maxima as IMERG.
EURADCLIM generally agrees with the overland precipitation accumulation spatial patterns of IMERG and E-OBS, with precipitation maxima magnitudes similar to E-OBS. EURADCLIM also provides additional insight into precipitation patterns over some coastal lands and their adjacent waters, allowing for comparisons to IMERG over the seas (Fig. 2c). EURADCLIM precipitation accumulation values exceed 2000 mm yr−1 along the coastal Norwegian waters, a larger value than observed in IMERG. Accumulations range 1000–1250 mm yr−1 along the western shores of Great Britain and Ireland and 750–1250 mm yr−1 over the Bay of Biscay coastal waters along northern Spain, accumulations which are lower than in IMERG. Accumulation values over the western shores of Denmark and the northwestern shores of Italy are more similar in magnitude to IMERG.
OPERA, the foundation for EURADCLIM, annually averaged accumulated precipitation (Fig. 2d) illustrates that the use of the radar-only OPERA product proves problematic for long-term climatological averages, with similar challenges encountered by others (e.g., M. Taszarek 2022, personal communication). In the development of the annual average, data gaps arise over large portions of Europe, including France, Belgium, Switzerland, Poland, Czechia, and Slovakia, not apparent in EURADCLIM, illustrating the advantages of the combined radar- and gauge-based product. Consequently, OPERA products are likely not suitable for climatological studies such as this one.
The spatial coverage of the ERA5, over land and sea, allows for a more direct comparison to the IMERG precipitation product, while providing relatively more detailed information over inland mountainous region for comparison to the E-OBS and EURADCLIM precipitation products (Fig. 2e). Over the sea, ERA5 precipitation accumulation values are smaller than IMERG over many regions, with maxima of 1000–1500 mm yr−1 over the northeastern Atlantic and northwestern coastal Spain into northern Portugal and 750–1000 mm yr−1 over the Adriatic coastal waters, the Ionian Sea, and the Mediterranean Sea along coastal Italy. In contrast, ERA5 precipitation accumulation over the Alps, Dinaric Alps, southwestern Balkan Peninsula, Carpathian Mountains, Pyrenees, and northern coastal Spain into northern Portugal are more expansive and 250–750 mm yr−1 larger than E-OBS, though smaller than E-OBS over the mountains of the Riviera del Levante of Italy and central Great Britain.
b. Annual convective precipitation accumulation patterns
While the above analysis highlights regions of enhanced precipitation over Europe and the surrounding seas, it includes all types of precipitation. As discussed in section 2c, intense convective precipitation is classified using a combination of IMERG precipitation rate with an imposed 10 mm h−1 threshold in conjunction with lightning activity in close spatiotemporal proximity to the heavy precipitation. Figure 3a highlights regions with the most intense rain rates, our first criteria for convective precipitation, represented as the annually averaged number of hours with rain rates ≥ 10 mm h−1. Intense rain rates occur most frequently, >36 h yr−1, over the northern and eastern Adriatic Sea, the coastal Ionia Sea along the southern Balkan Peninsula coast, and coastal Norway. Intense rain rates also occur > 24 h yr−1 farther offshore over the Ionian Sea, the Aegean Sea, the Mediterranean along the western Italian coastline, the Julian Prealps, Montenegro, coastal northwestern Spain, the west coast of Great Britain, and the west coast of Ireland. Values exceed 12 h yr−1 over parts of the Alps, the Dinaric Alps, the northeastern Atlantic, the coastal waters extending from the Baltic Sea along Norway westward to the Bay of Biscay along northern Spain, and the majority of the Mediterranean Sea. Thus, intense precipitation occurs most frequently along the European coasts and seas.
Plan view of annually averaged for the 10-yr period ranging 2011–20 (a) number of hours with the IMERG precipitation rate exceeding 10 mm h−1 (yr−1), (b) ATDnet lightning flash density (km2 yr−1), (c) convective precipitation accumulation, identified from IMERG+ATDnet (mm yr−1), (d) percent of accumulated precipitation that is classified as convective precipitation, identified from IMERG+ATDnet (% yr−1), and (e) number of hours with convective precipitation, identified from IMERG+ATDnet (h yr−1).
Citation: Monthly Weather Review 152, 7; 10.1175/MWR-D-23-0156.1
Comparison of regions with intense precipitation to the spatial pattern of lightning flash density provides an initial insight into the favored areas for convective precipitation over Europe and illustrates the components that comprise the definition of convective precipitation in this study. Annually, lighting is most frequent over the Ticino region, the Julian Prealps, and the Riviera del Levante of Italy, followed by the Pyrenees and the adjacent mountains in southeastern Spain, the Massif Central, western Italy and its coastal waters, the northern and eastern Adriatic Sea extending inland into the Dinaric Alps, the Ionian Sea extending into the southern Balkan Peninsula, and parts of the Carpathian Mountains (Fig. 3b). Thus, not all regions with frequent intense precipitation are associated with frequent lightning activity, and vice versa, indicating that regions with high lightning are not necessarily regions that experience frequent heavy convective precipitation and the associated flooding.
Indeed, the western Great Britain, western Ireland, and the northeastern Atlantic exhibit minimal convective precipitation annually, 0–20 mm yr−1 (Fig. 3c). The greatest convective precipitation accumulations, >300 mm yr−1, occur over the northern and eastern portions of the Adriatic Sea, the Ionian Sea, the coastal Aegean Sea, the Mediterranean along Italy’s western coast, and regions southeast of Spain and France. Other regions with relatively large accumulations, 150–250 mm yr−1, include portions of the Alps, the Julian Prealps, the western slope of the Dinaric Alps, coastal Montenegro and the southern Croatian Dalmatia region, the Massif Central, and southeastern Biscay Bay. Convective precipitation, with accumulations ranging 50–150 mm yr−1, also occurs over the waters south of Portugal, northwestern Spain and northern Portugal coastal regions, the Pyrenees, the English Channel extending eastward to the southern coastal waters of the North and Baltic Seas, southwestern coastal Norway, the Carpathians, and scattered regions across interior Europe. Thus, the largest convective precipitation accumulations are located over the European waters, particularly in the southern seas, and the mountains; there is a notable local minimum in convective precipitation over the tallest peaks of the Alps.
Given the flooding hazards associated with convective precipitation, it is beneficial to quantify the fraction of precipitation considered to be convective, highlighting regions that may be more vulnerable to these hazards. Based on our definition of convective precipitation, more than 30% of the annual accumulated precipitation over portions of the Ionian Sea is associated with convection and this is the largest percentage in the domain (Fig. 3d). Between 20% and 30% of precipitation is from convection over the much of the Mediterranean, Ionian, and Aegean Seas and the eastern Adriatic Sea into the coastal Balkan Peninsula. The remaining portions of the Mediterranean extending into the Strait of Gibraltar, the Massif Central, and the Julian Prealps receive 10%–20% of their precipitation from convection. Southeastern Biscay Bay, the Pyrenees, the remaining portions of the Massif Central, the Alps, the Carpathians, scattered portions of interior Europe, and the southern coastal waters of the North and Baltic Seas receive 5%–10% of their precipitation from convection.
Complementarily, the annually averaged number of hours with convective precipitation also provides insight into the vulnerability of regions to flooding based on the frequency of occurrence. Perhaps, as anticipated based on the above analysis, the eastern Adriatic and Ionian Seas experience the largest frequency of convective precipitation, more than 15 h yr−1 (Fig. 3e), while the Julian Prealps, coastal Montenegro, and portions of the Mediterranean Sea experience 7–13 h yr−1. This suggests that the large annual convective precipitation accumulations in these regions are, in part, due to the relatively large frequency of events. The remaining portions of the Mediterranean, Adriatic, Ionian Seas, Dinaric Alps, and Massif Central, as well as portions of the Alps, experience 5–7 h yr−1, while southeastern Biscay Bay, localized regions of Pyrenees and Carpathians, and the coastal waters of the North and Baltic Seas experience 3–5 h yr−1.
Corresponding analysis is performed to identify convective precipitation in the EURADCLIM dataset (Figs. 4a,b). Though EURADCLIM is unavailable over the European seas, Italy, and the southern Balkan Peninsula, regions with frequent convective precipitation and large accumulations, comparisons can be made with IMERG over the remaining land regions and coastal waters. Accumulations based on EURADCLIM are notably smaller than IMERG across the region. While IMERG indicates annual totals ranging 10–100 mm yr−1 across much of the continent (Fig. 3c), EURADCLIM values rarely exceed 50 mm yr−1, with many places receiving <10 mm yr−1 (Fig. 4a). Despite these magnitude differences, IMERG and EURADCLIM highlight similar locations as regions with relatively large annual totals, including the Alps extending southward into Po Valley and the northern Adriatic Sea, with localized maxima over the Italian Friuli Venezia Giulia region, the northwestern Balkan Peninsula, and Italy’s northwestern coastal waters; EURADCLIM accumulation totals range 50–100 mm yr−1 in these regions (Fig. 4a). As for IMERG, EURADCLIM indicates that these regions receive a larger percentage of annual precipitation from convection. Over the coastal Balkan Peninsula, as much as 20% of the annual precipitation is from convection, while values range 5%–10% over the Alps (Fig. 4b); similar percentage values are observed for IMERG (Fig. 3d).
Plan view of annually averaged for the 8-yr period ranging 2013–20 of EURADCLIM (a) convective precipitation accumulation, identified from EURADCLIM+ATDnet (mm yr−1), (b) percent of accumulated precipitation classified as convective, identified from EURADCLIM+ATDnet (% yr−1); plan view of the annually averaged for the 10-yr period ranging 2011–20 of ERA5 (c) convective precipitation accumulation (mm yr−1), (d) percent of precipitation produced by the convective scheme, based on the total precipitation accumulation (% yr−1), (e) convective precipitation accumulation for precipitation rates exceeding 10 mm h−1 (mm yr−1), (f) percent of convective precipitation accumulation from precipitation rates exceeding 10 mm h−1, based on the convective precipitation accumulation (% yr−1). Note the different shading ranges for (c) and (e) and (d) and (f).
Citation: Monthly Weather Review 152, 7; 10.1175/MWR-D-23-0156.1
c. Annual ERA5 convective precipitation accumulation patterns
Given the pervasive use of the ERA5 products, it is informative to evaluate the convective precipitation spatial patterns produced by the ECMWF IFS convective parameterization scheme in comparison with the observation-based convective precipitation defined in this study. The numerical model produces nearly twice as much convective precipitation across the domain, as compared to the IMERG (Fig. 4c, note the shading scale). Further, the spatial distribution of the greatest precipitation accumulations resembles those for the total annual precipitation from ERA5 and all observation-based products presented in Fig. 2, not the convective precipitation patterns presented in Figs. 3 and 4a,b. Thus, the fraction of precipitation generated by the convective scheme, versus the total precipitation, is notably high (Fig. 4d). For example, 50%–70% of precipitation over the Mediterranean, Ionian, and Adriatic Seas is produced by the numerical convective scheme. It is encouraging, however, that there is a relative decrease in the fraction of ERA5 precipitation produced by convection moving from southern to northern latitudes, as observed for the IMERG and EURADCLIM (Figs. 3d and 4b).
One cause of the large differences between the ERA5 and IMERG/EURADCLIM convective precipitation is the precipitation rate threshold, >10 mm h−1, imposed in our definition. Figure 4e illustrates the ERA5 accumulated convective precipitation from precipitation rates > 10 mm h−1 only (note that the shading interval and range are the same as in Figs. 3c and 4a for direct comparison). In contrast to all ERA5 convective precipitation, the annual accumulation of ERA5 convective precipitation for rates > 10 mm h−1 is notably lower than for IMERG over all European seas: While the IMERG indicates accumulations > 100 mm yr−1 across many of the northern coastal seas and >200 mm yr−1 across most of the southern seas, ERA5 produces <10 mm yr−1 over these regions. Over land, the differences between ERA5 and IMERG are less, though ERA5 values are smaller than IMERG in localized regions of large IMERG accumulations, including the Pyrenees, the Massif Central, the Alps, the Julian Prealps, and the southern Croatian Dalmatia region into Montenegro. ERA5 produces <10 mm yr−1 of convective precipitation over southwestern Norway and Denmark, though the IMERG indicates 50–100 mm yr−1. Accumulation values between ERA5 and EURADCLIM are more comparable, though ERA5 indicates widespread accumulations between 10 and 50 mm yr−1 over the continent, unlike EURADCLIM (Figs. 4c,e). Figure 4f illustrates the percent of ERA5 convective precipitation from precipitation rates > 10 mm h−1, versus all ERA5 convective precipitation. Less than 1% of ERA5 convective precipitation accumulations are from precipitation rates larger than 10 mm h−1 over most of Europe, with only a few localized regions over Italy, eastern Europe, and the southern Iberian coast receiving 1%–3%.
4. Seasonal precipitation patterns
a. Seasonal total precipitation accumulation patterns
Seasonal analyses of the different precipitation products illustrate the intra-annual variations within Europe’s annual total precipitation patterns. In spring (MAM), the largest IMERG precipitation accumulation maxima, >350 mm season−1, are located over coastal northwestern Spain and northern Portugal, the Pyrenees, the Massif Central, the Alps, the Dinaric Alps, and the northeastern Atlantic Ocean (Fig. 5a). Magnitude differences in seasonal precipitation accumulation across Europe, i.e., the domain-wide spatial gradients, are smallest during spring. In summer (JJA), notable differences exist across Europe (Fig. 5b). Localized areas of precipitation accumulations > 500 mm season−1 exist over the Alps and southwestern Norway, with increased activity over the northern Adriatic, the Carpathians, and the North and Baltic Seas. Low precipitation accumulations, < 75 mm season−1, encompass an expansive region over the southern latitudes, including the southern Mediterranean, Ionian, and Aegean Seas, and especially southern Spain and Portugal during the season. In autumn (SON), precipitation accumulation maxima shift to the European waters, with accumulations > 500 mm season−1 throughout the Mediterranean, Adriatic, and Ionian Seas, and the coastal Dinaric Alps (Fig. 5c). To the north, maxima also exist over western Ireland, western Great Britain, the northeastern Atlantic, and southwestern Norway and the adjacent coastal waters. Accumulation values over the Alps are lower in autumn as compared to summer, though there is a large increase to their southeast over northeastern Italy, Slovenia, Croatia, and Montenegro. Winter (DJF) precipitation accumulations remain high along the west coast of the Balkan Peninsula including Montenegro, the Dinaric Alps, the eastern Adriatic and Ionian Seas, and along the western Italy coastal waters, with an extensive maximum over the northeastern Atlantic Ocean extending southward to over northwestern Spain (Fig. 5d).
Plan view of seasonally averaged IMERG precipitation accumulation (mm season−1) for the 10-yr period ranging 2011–20 for (a) spring (MAM), (b) summer (JJA), (c) autumn (SON), and (d) winter (DJF).
Citation: Monthly Weather Review 152, 7; 10.1175/MWR-D-23-0156.1
Over land, seasonal patterns in E-OBS accumulated precipitation resemble those in IMERG, though with finer spatial gradients and thus localized larger values over the mountainous regions (Fig. 6).1 In spring, E-OBS exhibits precipitation maxima along the coastal mountains in the Riviera del Levante of Italy, western Ireland, western Great Britain, and southwestern Norway, the Italian Apennines, the Alps, the Pyrenees, the Massif Central, and the inland mountains of Germany, Poland, Czechia, and Great Britain (Fig. 6a). In summer, E-OBS illustrates the detailed complexity in the accumulation maxima seen in IMERG over the Pyrenees, the Massif Central, the Alps, the Carpathian Mountains, Norway, Great Britain, and Ireland (Fig. 6b). E-OBS also indicates a precipitation minimum over southern Spain and Portugal, as in IMERG. In autumn, E-OBS highlights the southeastern Massif Central, the Riviera del Levante, and Friuli Venezia Giulia in Italy, Slovenia, Croatia, Montenegro, and the western coasts of Norway, Ireland, and Great Britain as active regions (Fig. 6c), as in IMERG. In winter, distinct maxima are over northwestern Spain into northern Portugal, the Pyrenees, the Massif Central, Montenegro, the Riviera del Levante, the Julian Prealps, the Alps, and the interior mountains of Europe (Fig. 6d). In autumn and winter, western Norway, Great Britain, and Ireland continue to experience large accumulations.
As in Fig. 5, but for E-OBS precipitation.
Citation: Monthly Weather Review 152, 7; 10.1175/MWR-D-23-0156.1
EURADCLIM exhibits similar overland seasonal precipitation accumulation spatial patterns and magnitudes as IMERG and E-OBS (Fig. 7). The seasonal EURADCLIM precipitation accumulation patterns are somewhat noisier than the annual pattern, complicating the interpretation, in part, due to the reduced averaging period for these seasonal analyses as compared to the annual analysis, and noise in the OPERA product (Fig. 8), the basis of EURADCLIM. For example, there exists localized maxima radially around some of the OPERA radar locations (e.g., coastal Norway; Fig. 8c), interference signals that manifest as streaks of high precipitation accumulation extending radially away from a radar location (e.g., Mallorca of the Balearic Islands; Figs. 8a–d), and gaps in the data (e.g., France; Fig. 8d). Regardless, EURADCLIM highlights many important overwater precipitation features seen in IMERG, such as larger precipitation accumulation values over the northern Mediterranean in autumn (Fig. 7c) as compared to summer (Fig. 7b) and larger accumulations along the Norwegian coastal waters in autumn and winter (Figs. 7c,d). Over land, EURADCLIM also produces the large summertime maxima over the Alps (Fig. 7b) and maxima over northwestern Spain and northern Portugal in autumn and winter (Figs. 7c,d).
Plan view of seasonally averaged EURADCLIM precipitation accumulation (mm season−1) for the 8-yr period ranging 2013–20 for (a) spring (MAM), (b) summer (JJA), (c) autumn (SON), and (d) winter (DJF).
Citation: Monthly Weather Review 152, 7; 10.1175/MWR-D-23-0156.1
Plan view of seasonally averaged OPERA precipitation accumulation (mm season−1) for the 7-yr period ranging 2014–20 for (a) spring (MAM), (b) summer (JJA), (c) autumn (SON), and (d) winter (DJF).
Citation: Monthly Weather Review 152, 7; 10.1175/MWR-D-23-0156.1
b. Seasonal ERA5 precipitation accumulation patterns
Seasonal patterns in ERA5 total precipitation are also compared to IMERG with emphasis on coastal and offshore patterns and to E-OBS and EURADCLIM with a focus on inland patterns particularly over mountains (Fig. 9). While there are general similarities between ERA5 and IMERG accumulated precipitation patterns over the European seas, there are some notable differences. In summer, ERA5 exhibits comparatively small precipitation accumulations in the southern portion of the domain; values over the Mediterranean, Adriatic, and Ionian Seas are primarily < 50 mm season−1 in ERA5 (Fig. 9b), while IMERG values exceed 100 mm season−1. Only a few localized areas in the Mediterranean region receive > 75 mm season−1 in ERA5, specifically over the northern and eastern Adriatic and the waters southeast of Spain, which are active regions in IMERG as well during summer (Fig. 5b). In autumn, ERA5 seasonal accumulation values are several hundred millimeters lower than in IMERG over the southern seas and northeastern Atlantic (Figs. 5c and 9c). This difference appears in winter as well, with lower values in ERA5 over the northeastern Atlantic, the North Sea, the western Italian coastal waters, and the Balkan Peninsula coastal waters (Figs. 9d and 5d). ERA5 and IMERG overwater precipitation accumulation values are similar in spring (Figs. 5a and 9a).
Plan view of seasonally averaged ERA5 total precipitation accumulation (mm season−1) for the 10-yr period ranging 2011–20 for (a) spring (MAM), (b) summer (JJA), (c) autumn (SON), and (d) winter (DJF).
Citation: Monthly Weather Review 152, 7; 10.1175/MWR-D-23-0156.1
Conversely, ERA5 seasonal precipitation accumulation maxima values over land are generally spatially more expansive and larger in magnitude than E-OBS and EURADCLIM. In spring, this difference is most evident over northwestern Spain into northern Portugal, the Pyrenees, the Alps, the Apennines, the Dinaric Alps, and the Carpathian Mountains (Figs. 6a, 7a, and 9a). In summer, ERA5 values remain larger over the Pyrenees, the Alps, and the Carpathians (Figs. 6b, 7b, and 9b). In autumn, the biggest differences are over the Alps, western Italy, and the mountains along the southern Balkan Peninsula (Figs. 6c, 7c, and 9c), with ERA5 accumulations larger than the other products, though maxima over the Riviera del Levante of Italy, the Pyrenees, and the northern Dinaric Alps during autumn are similar among the products. In winter, precipitation values over land are more comparable among products, though ERA5 exhibits greater magnitude maxima over the west coast of the Balkan Peninsula, while E-OBS accumulations are larger over the Riviera del Levante (Figs. 6d and 9d). Accumulations over the southwest Norway, western Great Britain, and western Ireland coasts are comparable among products for all seasons.
c. Seasonal convective precipitation accumulation patterns
Toward an understanding of the seasonal spatial patterns of convective precipitation, Fig. 10 presents the average number of hours with IMERG precipitation rate ≥ 10 mm h−1 season−1. Intense precipitation is least frequent in spring, with 8–12 h of heavy precipitation only over southwestern Norway (Fig. 10a). Northwestern Great Britain, western Ireland, coastal northwestern Spain into northern Portugal, the Pyrenees, the Massif Central, the Alps, the coastal Adriatic, Montenegro, and parts of the Mediterranean experience 4–8 h of heavy precipitation. In summer, heavy precipitation is most common over the northern Adriatic Sea, with 8–12 h season−1 (Fig. 10b). Heavy precipitation also becomes more common over the Alps, North Sea, and Baltic Sea, with >4 h season−1. The frequency of intense precipitation is markedly larger in autumn, as compared to spring and summer, with the largest activity over the Mediterranean region (Fig. 10c). Over the Ionia Sea, intense precipitation occurs more than 12 h season−1, while over much of the Mediterranean, coastal Adriatic Sea, Julian Prealps, and coastal Montenegro/Dalmatia regions, it occurs 8–12 h season−1. Intense precipitation is also common over the remaining Dinaric Alps, the Alps, the Massif Central, northwestern Spain, western Ireland, western Great Britain, and southwestern Norway. In winter, the frequency of heavy precipitation maximizes over the northwestern part of the domain, though there are still maxima along the coastal Balkan Peninsula, northern Adriatic, and western Italian coast (Fig. 10d).
Plan view of the number of hours with IMERG precipitation rates exceeding 10 mm h−1 (h season−1) for the 10-yr period ranging 2011–2020 for (a) spring (MAM), (b) summer (JJA), (c) autumn (SON), and (d) winter (DJF).3
Citation: Monthly Weather Review 152, 7; 10.1175/MWR-D-23-0156.1
Lightning activity, a requirement for our convective precipitation classification, is greatest over land during summer and over water during autumn (Figs. 11b,c). In summer, maxima in activity are collocated with the elevated terrain in many regions, including the coastal mountains of northeastern Spain, the Pyrenees, the Massif Central, the Alps, the Apennines, the Dinaric Alps, and the Carpathian Mountains (Fig. 11b). Over the sea, lightning is most frequent over the northern Adriatic, with enhanced activity over the remaining portions of the Adriatic Sea, the Ionian and Aegean Seas, and the eastern Mediterranean along the Italian coast. There is also generally widespread activity over much of the European continent south of the North and Baltic Seas and east of central Spain. In autumn, there is a dramatic southward shift in the lightning density maxima, with the greatest activity confined to the southern seas and the adjacent coasts (Fig. 11c). Within this region, the greatest lightning flash density occurs over the Ionian Sea, the eastern Adriatic and western Balkan coast, and the west coast of Italy. In winter, lightning activity is confined to the southern seas, with a relative maximum over the eastern Ionian Sea (Fig. 11d). In spring, there are relative maxima over the Alps, the Pyrenees, and portions of the Balkan Peninsula, with widespread weaker activity over the remaining European land regions (Fig. 11a).
Plan view of seasonally averaged ATDnet lightning flash density (flashes km−2 season−1) for the 10-yr period ranging 2011–20 for (a) spring (MAM), (b) summer (JJA), (c) autumn (SON), and (d) winter (DJF).
Citation: Monthly Weather Review 152, 7; 10.1175/MWR-D-23-0156.1
Based on the spatial patterns in the frequency of heavy precipitation and lighting activity, one may anticipate that heavy convective precipitation peaks over land in summer and over the sea in autumn. Indeed, convective precipitation in summer is widespread across the European continent east and north of Spain as well as the surrounding seas (Fig. 12b). In fact, though convective precipitation accumulations are generally larger over land, precipitation accumulations are largest, >50 mm season−1, over the northern Adriatic Sea, while the coastal North and Baltic Seas and across Denmark receive 40–50 mm season−1. Over land, accumulations are largest over the Alps, with peak values comparable to those over the northern coastal seas. Autumn features the most expansive region of the largest accumulations. Convective precipitation primarily occurs over the Mediterranean region, with accumulations > 70 mm season−1 over the Ionian Sea and widespread accumulations 50–70 mm season−1 over the Mediterranean and Adriatic Seas (Fig. 12c). Convective precipitation accumulations are also relatively large over Italy, the Pyrenees, the Massif Central, and western Balkan coast, particularly coastal Montenegro and the Julian Prealps. Over the remaining inland areas, there exists a distinct minimum in convective precipitation. In winter, convective precipitation accumulations remain large along the western Balkan coast and the eastern Adriatic and Ionian Seas, with elevated values over the Mediterranean Sea into the west coast of Italy, northwestern Spain in northern Portugal, and the coastal Biscay Bay (Fig. 12d). Convective precipitation accumulation maxima magnitudes are smallest in spring, with only 30 mm season−1 over the Pyrenees, the Alps, the western Balkan Peninsula, and the eastern Adriatic Sea (Fig. 12a).
Plan view of seasonally averaged IMERG heavy convective precipitation accumulation (mm season−1) for the 10-yr period ranging 2011–20 for (a) spring (MAM), (b) summer (JJA), (c) autumn (SON), and (d) winter (DJF).
Citation: Monthly Weather Review 152, 7; 10.1175/MWR-D-23-0156.1
Maxima in convective precipitation frequency (Fig. 13) are collocated with those for the precipitation accumulations, indicating that the number of precipitation events, in part, is responsible for the large local accumulation values. On average, springtime convective precipitation occurs only 1–3 h season−1 over the Alps, western Balkan coast, and eastern Adriatic and Ionian Seas (Fig. 13a), contributing to the small magnitude maxima in precipitation accumulations (Fig. 12a). In summer, frequency peaks over the northern Adriatic, with 6–9 h season−1 of convective precipitation, resulting in a local maximum in convective precipitation accumulation (Figs. 12b and 13b). Convective precipitation is also more frequent over the Julian Prealps, the coastal Aegean Sea, parts of the Alps, and the southern Baltic Sea. Autumn boasts the highest frequency of convective precipitation (Fig. 13c). Precipitation is most frequent over the Ionian Sea, with peak values of 15–18 h season−1, contributing to the regional peak in accumulations (Figs. 12c and 13c). Convective precipitation is also frequent over adjacent seas, with large portions of the Mediterranean experiencing 6–9 h season−1 of convective precipitation. In winter, convective precipitation is most frequent over the western Balkan coast, the eastern Adriatic and Ionian Seas, and the west coast of Italy, collocated with the accumulation maxima (Figs. 12d and 13d). Some portions of the Mediterranean and coastal Spain receive their wintertime convective accumulations from <3 h of precipitation, while other portions experience <1 h of precipitation, suggesting that relatively fewer intense events may be the cause of these maxima.
Plan view of the number of hours with heavy convective precipitation (h season−1) in IMERG for the 10-yr period ranging 2011–20 for (a) spring (MAM), (b) summer (JJA), (c) autumn (SON), and (d) winter (DJF).
Citation: Monthly Weather Review 152, 7; 10.1175/MWR-D-23-0156.1
Though the EURADCLIM annual, and therefore seasonal, convective precipitation accumulations are appreciably lower than IMERG, the two products exhibit similar locations in seasonal maxima and the distinct shift in convective precipitation from land to sea during the transition from summer to autumn (Fig. 14). Specifically, EURADCLIM exhibits widespread convective precipitation over the European continent in summer, with maxima over the northwestern Balkan Peninsula and northeastern Italy (Fig. 14b), a signal produced by IMERG as well, increasing confidence that this localized Italian maximum is robust. In autumn, EURADCLIM precipitation maxima are confined to the south over the Massif Central, northern Italy including Po Valley, the western Balkan coast, and portions of the northern and western Mediterranean (Fig. 14c). Though there is reduced availability of EURADCLIM in the southern portion of the domain, the locations of these existing maxima agree with the autumn precipitation maxima in IMERG (Fig. 12c), again indicating the robustness of this signal. Since most of the wintertime convective precipitation occurs outside the EURADCLIM domain according to IMERG, there is little convective precipitation in the EURADCLIM winter analysis (Fig. 14d). In spring, convective precipitation activity increases over the continent (Fig. 14a), as in IMERG.
Plan view of seasonally averaged EURADCLIM heavy convective precipitation accumulation (mm season−1) for the 8-yr period ranging 2013–20 for (a) spring (MAM), (b) summer (JJA), (c) autumn (SON), and (d) winter (DJF).
Citation: Monthly Weather Review 152, 7; 10.1175/MWR-D-23-0156.1
d. Seasonal ERA5 convective precipitation accumulation patterns
An evaluation of ERA5 seasonal convective precipitation highlights the high accumulation totals (Fig. 15; note the shading scale is as in Figs. 5–9) as compared to IMERG and EURADCLIM accumulations. In all seasons, ERA5 has large convective precipitation accumulations over the Atlantic, as large as 250 mm season−1 in winter, though IMERG indicates <10 mm season−1 (Figs. 12 and 15). In summer, ERA5 identifies the Pyrenees, Alps, and Carpathians as regions with the greatest accumulations; however, IMERG indicates that its accumulation magnitudes over these mountains are similar to or lower than accumulations over the Adriatic, Mediterranean, North, and Baltic Seas (Figs. 12b and 15b).2 In autumn, ERA5 has maxima over the western Balkan Peninsula, western Italy, and their adjacent coastal waters. While IMERG has maxima in these locations as well, it also shows similarly large accumulations across the Mediterranean Sea and even higher accumulations over the Ionian Sea (Figs. 12c and 15c). In winter, ERA5 indicates widespread convective precipitation over the European continent, unlike IMERG and EURADCLIM which show convective precipitation confined to the southern latitudes of the domain (Figs. 12d, 14d and 15d).
Plan view of seasonally averaged ERA5 convective precipitation accumulation (mm season−1) for the 10-yr period ranging 2011–20 for (a) spring (MAM), (b) summer (JJA), (c) autumn (SON), and (d) winter (DJF).
Citation: Monthly Weather Review 152, 7; 10.1175/MWR-D-23-0156.1
Considering ERA5 convective precipitation rates > 10 mm h−1 only results in more agreement between ERA5, IMERG, and EURADCLIM (Fig. 16). ERA5 heavy convective precipitation accumulations are highest over the land during summer, with maxima > 40 mm season−1 over the Alps (Fig. 16b), consistent with the magnitude in IMERG. There remain, however, some distinctive differences between the products as well. ERA5 shows similarly large precipitation accumulations over the Pyrenees, the Carpathians, and Italy, unlike IMERG. Though ERA5 indicates summertime heavy convective precipitation accumulations over the northern Adriatic, values are 30–40 mm season−1 smaller than in IMERG (Figs. 12b and 16b). Further, ERA5 has no summertime heavy convective precipitation over the Mediterranean, southern Adriatic, Ionian, and Aegean Seas nor the North and Baltic Seas. In autumn, ERA5 heavy convective precipitation maximizes over the southern domain, as in IMERG and EURADCLIM, though accumulations are confined to the coastal regions; 20–30 mm season−1 accumulates along the southern Balkan and western Italian Peninsulas (Fig. 16c). Unlike IMERG and EURADCLIM, little convective precipitation occurs over the southern seas in ERA5. Stark differences arise in winter, with ERA5 producing no heavy convective precipitation (Fig. 16d), unlike IMERG and EURADCLIM. In spring, though the overall magnitudes of ERA5 heavy convective precipitation accumulation maxima are similar to IMERG, accumulations are less widespread in ERA5, with only isolated maxima along the coastal Balkan Peninsula, southern Italy, the Pyrenees, and Portugal (Figs. 12a and 16a). ERA5 has no heavy convective precipitation over the Alps, a local maximum in IMERG, nor over the southern seas. In contrast, the magnitude and spatial coverage of precipitation accumulations in ERA5 and EURADCLIM are generally similar in spring. Overall, much of the annual ERA5 heavy convective precipitation is due to summertime convection, with a small secondary contribution in autumn over portions of the domain southern latitudes.
As in Fig. 15, but for ERA5 convective precipitation accumulated from convective precipitation rates > 10 mm h−1 only.
Citation: Monthly Weather Review 152, 7; 10.1175/MWR-D-23-0156.1
5. Regional precipitation
a. Regional convective precipitation annual cycles
Based on the annual and seasonal analyses, we have selected eight subregions for further analysis due to their enhanced convective precipitation activity during the year: the Alps, northwestern Spain and northern Portugal, Po Valley, the Pyrenees Mountains, Italy, the southern Balkan Peninsula, and the northern and southern Adriatic Sea (Fig. 17). All regions display unique convective precipitation patterns, partly due to their differing landscapes, which include inland and/or coastal mountains with a wide range in relief and regions of only seas (Fig. 1a). The inland Alps have the tallest mountain peaks in the domain, >4500 m in altitude, while the adjacent plains of Po Valley are only 0–550 m above sea level. The Pyrenees, a ∼500-km-long mountain range located between Spain and France, have peaks > 3000 m; though the center of the mountain chain is located inland, the chain extends to the northern and southern coasts. Northwestern Spain and northern Portugal have expansive mountains ranging 500–1500 m in altitude adjacent to the Atlantic Ocean. The Italy regional box encompasses coastlines and the Apennine Mountains with peaks up to ∼2900 m in elevation, and the southern Balkan Peninsula is a mountainous coastal region with peaks ∼2300 m. Finally, the Adriatic Sea subregion is primarily marine, though includes the bounding coastal land, and is divided into two separate regions (North and South Adriatic) due to latitudinal variations in their convective precipitation signals. Though we limit our analysis to these eight regions, we acknowledge that other European subregions exhibit locally enhanced convective precipitation activity during the year, for example, the Ionian and Mediterranean Seas, the southern North and Baltic Seas, and the Massif Central mountains.
The eight subdomains chosen for the regional analysis: the Alps (dark orange), northwestern Spain and northern Portugal (light orange), the Pyrenees (yellow), Po Valley (gray), Italy (green), the northern Adriatic Sea (light blue), the southern Adriatic Sea (dark blue), and the southern Balkan Peninsula (purple).
Citation: Monthly Weather Review 152, 7; 10.1175/MWR-D-23-0156.1
The annual distribution of convective precipitation is analyzed for each subregion through time series of domain-averaged daily convective precipitation rates (Fig. 18). Values are normalized by the size of the domain and a 30-day moving average is applied to daily time series data to reduce noise. Enhanced convective precipitation rates over the inland Alps span much of the warm season, May–September, with a minimum in January–March. Of the subdomains analyzed in this study, this is the only region that experiences a clear summertime maximum in convective precipitation. Over the Pyrenees, the annual cycle in convective precipitation rates is less defined, exhibiting a gradual increase in convective precipitation beginning in early spring, a peak in October, and a relatively rapid subsequent decline in November–December. Over the coastal mountains of northwestern Spain and northern Portugal, convective precipitation activity maximizes during autumn into winter, with a secondary maximum in late spring to early summer and minimum in convective precipitation during July–August.
Annual cycle of area-averaged regional convective precipitation rates (mm day−1) for each of the eight subdomains (see Fig. 17), including the Alps (dark orange), northwestern Spain and northern Portugal (light orange), Po Valley (gray), the Pyrenees (yellow), Italy (green), the northern Adriatic Sea (light blue), the southern Adriatic Sea (dark blue), and the southern Balkan Peninsula (purple).
Citation: Monthly Weather Review 152, 7; 10.1175/MWR-D-23-0156.1
Over the Mediterranean land regions, the peak in convective precipitation is predominantly during the later autumn months and into winter. Over Po Valley, convective precipitation rates increase gradually beginning in April, with a peak in activity in October–November, followed by a dramatic decrease in winter. The annual cycles of convective precipitation over the Italian and southern Balkan coastal mountain regions parallel one another. Over Italy, convective precipitation markedly increases in August and peaks in November–December, with a relative minimum from March to July. Over the southern Balkan Peninsula, convective activity increases in September and peaks in November–January, with a relative minimum between April and August. The maxima in area-averaged convective precipitation rates for these eastern Mediterranean domains are larger than for the three mountainous regions discussed above.
Over the northern and southern Adriatic Sea, the magnitudes of convective precipitation rates are substantially larger than all other subregions, twice as large as rates over Italy and the Balkans, while 3–4 times as large as rates over the Alps. Over the northern Adriatic, convective precipitation activity increases in May–June and peaks in July–October with a secondary annual maximum in November–December. Over the southern Adriatic, the peak activity shifts slightly later in the year, with a more gradual increase from May to July, a peak in August–January, and a rapid decrease into spring. Thus, convective precipitation activity over the northern Adriatic Sea is approximately 1 month earlier than over the southern Adriatic. Considering the annual cycle of convective precipitation over all subdomains, the end of March into early April is a conspicuously inactive time for convective precipitation.
b. Regional convective precipitation diurnal cycles
The monthly averaged diurnal cycle in convective precipitation rates is evaluated for each subregion to quantify the time of day during which regional convective precipitation is most intense during the year and more importantly during the most active months in the year (Fig. 19). Convective precipitation rate data are area and hourly averaged for each month, and a b-spline interpolation is applied to each monthly diurnal cycle. Convective precipitation over the Alps is favored during summer from May to September and precipitation rates follow a well-defined classic diurnal cycle during these months, with minima in the morning and maxima in the late afternoon to the early evening (Figs. 19e–h, red). Precipitation intensity increases in the morning after 0900 UTC and peaks in the late afternoon between 1600 and 1900 UTC, though peak intensity occurs later in the afternoon in late summer; in August, convective precipitation rates maximize between 1800 and 2100 UTC. Over the Pyrenees, during the broad secondary summertime peak in convective precipitation, there is a less pronounced though similar diurnal cycle, with precipitation rates maximizing in the afternoon to the early evening and declining in the overnight to morning hours (Figs. 19e–k, yellow). The peak in convective activity over northwestern Spain and northern Portugal occurs during the late overnight through the morning (0600–1100 UTC), with this pattern most apparent in November, the most active month for convective precipitation (Fig. 19k, orange). Over Po Valley, there are distinct differences in the diurnal cycle of convective precipitation during the most active months of October and November. In October, convective precipitation peaks in the early afternoon (1100–1300 UTC), with a secondary peak during the late evening (2000–2200 UTC; Fig. 19j). In November, convective precipitation peaks in the early part of the morning (0500–0700 UTC; Fig. 19k).
Diurnal cycle of area-averaged regional convective precipitation rates (mm h−1) for each of the eight subdomains (see Fig. 17) including the Alps (dark orange), northwestern Spain and northern Portugal (light orange), Po Valley (gray), the Pyrenees (yellow), Italy (green), the northern Adriatic Sea (light blue), the southern Adriatic Sea (dark blue), and the southern Balkan Peninsula (purple) in (a) January, (b) February, (c) March, (d) April, (e) May, (f) June, (g) July, (h) August, (i) September, (j) October, (k) November, and (l) December.
Citation: Monthly Weather Review 152, 7; 10.1175/MWR-D-23-0156.1
Diurnal fluctuations in convective precipitation over the remaining four subregions are less clear compared to the regions discussed above. Over Italy, during the convectively active months of November–December, there is essentially no evidence of a diurnal cycle (Figs. 19k,l, green), indicating that convective precipitation is equally likely throughout the day. Similarly, over the southern Balkan Peninsula, there is no obvious diurnal cycle in precipitation during the most active months of November–January (Figs. 19k,l, purple). Over the northern Adriatic Sea during late summer, there is evidence of a diurnal variation in convective precipitation, with a peak in the late afternoon and evening (1600–2100 UTC) during July and August (Figs. 19g,h, light blue). This signal is not evident during autumn, with no apparent diurnal cycle during these most active months (Figs. 19i–k). Over the southern Adriatic Sea, there is little evidence of a diurnal cycle in the convective precipitation rate during any month of the year (Fig. 19, dark blue).
6. Discussion
This study assessed the annual, seasonal, and diurnal intense convective precipitation patterns across a pan-European domain using a combination of precipitation and lightning products. A dominant signal shown is the prolific heavy convective precipitation activity over the Mediterranean in autumn (Figs. 12c–14c), which agrees with previous studies (Mehta and Yang 2008; Alhammoud et al. 2014; Taszarek et al. 2019a, 2020a). One contributing factor is the relatively high Mediterranean sea surface temperatures (SSTs) during the season (Fig. S1 in the online supplemental material). Kotroni and Lagouvardos (2016) showed that sensible and latent heat fluxes associated with high SSTs increase the lower-tropospheric temperature and steepen the environmental lapse rate. This, combined with the large moisture availability from the underlying sea, can lead to locally enhanced convective activity. In their lightning climatology, Enno et al. (2020) hypothesized that the Mediterranean autumnal lightning maximum is related to cyclonic storms advecting cool continental air from Europe over the warm waters of the Mediterranean, creating a conditionally unstable environment conducive for convection. Once over the land surface, these storms dissipate, thus explaining the precipitation maxima over water and the Mediterranean coastlines and the relative minimum over inland areas (Figs. 12c–14c). Similarly, based on soundings, surface observations, European Cooperation for Lightning Detection (EUCLID) and ZEUS lightning data (Anagnostou et al. 2002), ERA-Interim, and severe weather reports, Taszarek et al. (2019a) illustrated a peak in thunderstorm days along the coastal zone of the Mediterranean Sea in autumn.
Though SSTs are high during July and August (Fig. S1), precipitation over the Mediterranean is less frequent than over land (Figs. 12b and 13b). This is, in part, due to the latitudinal shift in the location of the Azores high throughout the year and the resulting pressure modification over the Mediterranean region. In summer, the Azores high creates a prominent blocking pattern over the region and the associated subsidence inhibits convection, though this high pressure moves northward away from the Mediterranean in autumn (Sanders 1953; Rashid et al. 2012). Consequently, there is greater lightning activity and heavy convective precipitation over land and less over the sea during summer (Figs. 11b–13b). Additionally, Anderson and Klugmann (2014) hypothesized that the primary mechanism for summer convection is diurnal heating, which is consistent with the diurnal pattern in convective precipitation observed over the Alps in this study (Figs. 19e–h). Taszarek et al. (2019a) also found that June–July has the most thunderstorm days over inland regions, with maxima over the western and eastern Mediterranean in autumn and winter, respectively. They hypothesized that the summer lightning minimum over water is due to the stabilizing effect of the relatively cold waters, though this is likely true during the early summer as regional SSTs maximize in August (Fig. S1).
Considering the regional subdomains explored in this study, the observed seasonal and diurnal patterns of Alpine convective precipitation are largely consistent with prior studies. Analyzing high-resolution rain gauge data, both Frei and Schär (1998) and Isotta et al. (2014) described a precipitation maximum extending along the northern mountain rim, a feature seen in our summer-averaged analysis of all precipitation (Figs. 6b–9b) and convective precipitation (Figs. 12b and 16b). Taszarek et al. (2019a) also showed that thunderstorms are more frequent over the Alps in summer, as compared to other regions in Europe. Isotta et al. (2014) found summer precipitation in the Alps to be nearly twice as large as in winter. Enhanced diurnal heating and orographic upslope likely initiate convection over the mountainous regions in summer.
In the Alpine lee, Po Valley is identified as a region with enhanced convective activity (Fig. 18) and has been an actively studied region for convective precipitation through climatologies, observational case studies, and numerical simulations (Paccagnella et al. 1992; Buzzi and Alberoni 1992; Cacciamani et al. 1995; van Delden 2001; Reeves and Lin 2006; Collino et al. 2009; Feudale and Manzato 2014; Comin et al. 2015; Pucillo et al. 2020). Based on 18 years of EUCLID, Feudale and Manzato (2014) illustrated a peak in lightning activity over low-lying portions of Po Valley from August to September, consistent with our secondary annual peak in convective precipitation (Fig. 18), and greater activity during the evening and overnight of these months, consistent with our diurnal cycle of convective precipitation during these same months (Figs. 19h,i). Over the Prealps, they found a peak in lighting activity in July with a maximum between 1600 and 1700 UTC (Feudale and Manzato 2014), as for convective precipitation over our Alps domain (Figs. 18 and 19g). Research has emphasized the important role of synoptic-scale dynamic forcing, particularly associated with Alpine lee cyclogenesis and the passage of fronts, in the development and intensification of convective precipitation across Po Valley. For example, van Delden (2001) asserted that from April to October, thunderstorms south of the Alps are initiated by lee cyclones and sea breezes, based on analyses of case studies. In the presence of a surface-based leeside stable layer, the location of convection in the valley is sensitive to the density of the layer, with convection forming further south for stronger stable layers (Reeves and Lin 2006).
Further south over the Italian peninsula, our results highlight autumn and early winter as the period with the most convective precipitation (Fig. 18), which contrasts with the June peak in the number of thunderstorm days identified by Taszarek et al. (2019a). Such differences may indicate that the environment is more often favorable for convection in summer based on their defined metrics, but the associated precipitation is more intense in autumn. Regardless, both studies identify Italy as a particularly active region for convective precipitation. Comin et al. (2015) argued for the importance of sea-breeze convergence as a lifting mechanism in the Salento Peninsula in southeastern Italy and its role in increasing low-level moisture over land, increasing CAPE, and reducing convective inhibition (CIN).
Both this study and Taszarek et al. (2019a) also identify the southern Balkan Peninsula as a convectively active area in winter. Kotroni and Lagouvardos (2016) explain that heat fluxes from high SSTs destabilize the atmosphere by warming the near-surface air, enhancing convection over the southern Balkan Peninsula. Recently, anomalous high SSTs contributed to the development of a long-lived, organized convective storm that produced a swath of severe winds from the Balearic Islands eastward into Czechia (González-Alemán et al. 2023). For the adjacent Adriatic Sea, convective precipitation rates over the northern Adriatic increase ∼1 month earlier in the year than the southern Adriatic (Fig. 18), implying that precipitation along the long axis of the sea may be affected by the summer blocking high pressure; the later precipitation maximum over the southern Adriatic Sea may be due to the pressure maximum’s closer proximity to the central Mediterranean (Sanders 1953; Rashid et al. 2012). Convective precipitation over the northern and southern regions of the Adriatic Sea exhibits little diurnal cycle during the most active months (Fig. 19). A similar behavior was observed over the expanse of the Mediterranean Sea by Alhammoud et al. (2014). The weak diurnal cycle of convective precipitation observed during July–August over the northern Adriatic may be a reflection of convection originating over land during the peak heating moving over the northern extent of the sea.
A number of studies have examined convective precipitation over Spain’s Catalonia region, which extends northward toward the southern Pyrenees (e.g., Llasat and Puigcerver 1997; Jansa et al. 2014; Bech et al. 2015; Llasat et al. 2016; del Moral et al. 2017; Llasat et al. 2021). Based on rainfall rates obtained from precipitation gauges, Llasat et al. (2021) found that as much as 16% of the regional accumulated precipitation is from convection, with the highest annual totals near the pre-Pyrenees, and identified summer as the most active period for convective precipitation. Our work suggests that only ∼5% of all precipitation in the region is from convection (Fig. 3d), with the peak in activity during autumn (Fig. 18). These differences are likely a consequence of the different domains (western versus eastern Spain) and datasets analyzed and the different methods used to identify convective precipitation in our studies. Complementary to our findings, Taszarek et al. (2019a) showed a peak in thunderstorm activity over the coastal regions of Spain and Portugal in October. Regardless, studies emphasize the important role of the orography in supporting the development of convective precipitation in the region through the orographic ascent of moist air from the Mediterranean and the development of localized pressure and convergence features (Llasat and Puigcerver 1997; Jansa et al. 2014; del Moral et al. 2017). Further, Trapero et al. (2013) showed that the orographic uplift of moist flows from the Mediterranean was primarily responsible for heavy precipitation during an October and a November case study event, with the flow primarily driven by synoptic-scale processes. Thus, the maximum in convective precipitation during October–November found in our study (Fig. 18) may be, in part, due to synoptic-scale flow patterns.
Da Silva and Haerter (2023) developed a pan-European climatology of mesoscale convective systems (MCS), utilizing a combination of lighting products and IMERG precipitation rates, as in our study. Though their MCS study focused on one specific convection organizational mode, our findings are consistent, likely due to our usage of the same precipitation product. For example, MCS activity is greatest over the continent in summer and over the adjacent seas and ocean in the autumn and winter, a signal observed for all convective precipitation (Fig. 12). They argue for the predominant role of the diurnal cycle in controlling MCS activity over land in summer and the importance of both dynamics and thermodynamics for storms over the coast. They also identify northwestern Spain in October–November, the Adriatic and Ionian Seas in September–December, and the Alps in July–August as regions with frequent MCSs, consistent with our annual analyses. One notable difference from our findings is the percent of precipitation attributed to MCSs during a season. Da Silva and Haerter (2023) indicate that MCSs contribute to a large fraction of the total annual precipitation, contributing as much as 50%–80% in autumn over many regions, though a comparison of our Figs. 5c and 12c yields at most 10%–15%. These contrasting results may be a consequence of the different precipitation rate thresholds imposed in our studies during the identification of convective precipitation, 2 mm h−1 in Da Silva and Haerter (2023) versus 10 mm h−1 in this study. Analogously, we found large contributions from convective precipitation to the total annual sum in ERA5 (Figs. 4d and 15) until we imposed a 10-mm h−1 precipitation rate threshold (Fig. 16).
7. Conclusions
A 10-yr climatology of heavy convective precipitation over Europe was developed using multiple datasets, including satellite, rain gauge, ground-based radar, and reanalysis precipitation products in conjunction with lightning data. Specifically, annual and seasonal patterns of intense convective precipitation were quantified over the pan-European domain, and additional diurnal analyses were performed for eight subdomains. Convective precipitation was identified in IMERG and EURADCLIM, and requirements include precipitation rates exceeding 10 mm h−1 with lighting activity within 20 km and ± 30 min of this heavy precipitation point. Primary findings include the following:
-
Annually, regions with the largest convective precipitation accumulations are the European waters, specifically the northern and eastern portions of the Adriatic Sea, the Ionian Sea, the coastal Aegean Sea, and the Mediterranean along Italy’s western coast and southeast of Spain and France.
-
Over the European continent, up to 10% of the total annual precipitation accumulated is classified as convective precipitation, increasing to as large as 40% over the surrounding seas. In contrast, up to 50% of ERA5 annual precipitation over land is produced by the convective parameterization scheme increasing to as large as 70% over the seas; however, only 1% of ERA5 convective precipitation annual accumulations are from rates exceeding 10 mm h−1.
-
Heavy convective precipitation in summer is most common over continental Europe, and seasonal accumulations over the European waters are generally smaller during this time, except for the northern Adriatic Sea and the coastal North and Baltic Seas (extending across Denmark). In autumn, these maxima shift southward to the Mediterranean, Adriatic, and Ionian Seas and their coasts. Over land, accumulations are also relatively large over western Italy, the Pyrenees, the Massif Central, and the western Balkan Peninsula coast in autumn. Convective precipitation is less common in winter and spring. During winter, convective precipitation maximizes along the coastal Balkan Peninsula and the eastern Adriatic and Ionian Seas, with activity also relatively high over western Italy and its adjacent coastal waters, northwestern Spain into northern Portugal, and the majority of the Mediterranean Sea. In spring, convective precipitation predominately occurs over the Alps, the Pyrenees, the coastal Balkan Peninsula, and the eastern Adriatic Sea.
-
Regionally averaged convective precipitation rates over the Alps and Pyrenees Mountains in summer exhibit a well-defined diurnal cycle with a peak in mid- to late afternoon. Convective precipitation rates over the Alps are largest during this season. From autumn into early winter, regional convective precipitation rates peak over many of the analyzed subdomains, including northwestern Spain/northern Portugal, Po Valley, Italy, the southern Balkan Peninsula, and the northern and southern Adriatic Sea. Consequently, convective precipitation rates over northwestern Spain and northern Portugal are largest during the late overnight through morning, rather than during the daytime. Over Italy, the southern Balkan Peninsula, and the Adriatic Sea, there is little evidence of a favored time of day for convective precipitation. Po Valley has differing diurnal signals during the most convectively active months of October and November. In October, convective precipitation rates peak in the early afternoon, while in November, rates maximize in the early part of the morning.
Results from this study provide insight into the dominant annual and seasonal spatial patterns in heavy convective precipitation and illustrate the favored time of day for the most intense convective precipitation over a subset of convectively active regions. Thus, this work may be leveraged for situational awareness for weather forecasting applications, ideally promoting improved prediction of hazardous flooding. Additionally, this information may be useful for climate risk assessments, establishing a baseline for current convective precipitation patterns and aiding in anticipating future changes as the climate warms (Brooks 2013; Allen 2018). We acknowledge, however, that numerous scientific questions remain regarding pan-European convective precipitation. For example, future work can evaluate the dominant mesoscale processes supporting the spatiotemporal convective precipitation patterns observed in this study to reveal the primary and most common underlying processes responsible for the annual, seasonal, and regional diurnal cycles. Additional work may also leverage the IMERG satellite product to quantify the dominant initiation locations of convective elements (Manzato et al. 2022), their common tracks across the continent, and their decay locations toward improved predictability of hazardous flooding associated with these storms.
Acknowledgments.
This research was supported by the National Science Foundation Award AGS-2002660. The authors deeply thank colleagues at the European Severe Storms Laboratory (ESSL) and those in the European Severe Storms Community, including Pieter Groenemeijer, Alois Holzer, Francesco Battaglioli, Homa Ghasemifard, Mateusz Taszarek, Melita Perčec Tadić, Ksenija Cindric, and many others for the insightful discussions through the development of this manuscript. The authors are grateful to the Met Office for maintaining and supplying the ATDnet lightning data used in this work, as well as Météo-France for their assistance in obtaining and accessing the OPERA data. The authors also thank three anonymous reviewers for their constructive comments.
Data availability statement.
IMERG precipitation rate data are openly available from the National Aeronautics and Space Administration (NASA) Global Precipitation Measurement web portal at https://gpm.nasa.gov/data/imerg. E-OBS and ERA5 precipitation products are openly available through the Copernicus Climate Change Service C3S Climate Data Store (CDS; Buontempo et al. 2020) implemented by the European Centre for Medium-Range Weather Forecasts (ECMWF; https://cds.climate.copernicus.eu/cdsapp#!/dataset/reanalysis-era5-single-levels?tab=overview). EURADCLIM 1-h precipitation accumulation data are freely accessible from the Koninklijk Nederlands Meteorologisch Instituut (KNMI) Data Platform (https://dataplatform.knmi.nl/dataset/rad-opera-24h-rainfall-accumulation-euradclim-1-0). Due to confidentiality agreements, OPERA radar data are available from the Meteo-France Odyssey web portal (https://portail-api.meteofrance.fr/devportal/apis) for research and education usage only subject to a nondisclosure agreement. Details of the data and how to request access are available from EUMETNET (https://www.eumetnet.eu/activities/observations-programme/current-activities/opera/). OPERA radar locations are freely available from the European National Meteorological Services (EUMETNET; https://www.eumetnet.eu/wp-content/themes/aeron-child/observations-programme/current-activities/opera/database/OPERA_Database/index.html). Due to its propriety nature, ATDnet lightning data cannot be made publicly available. For data inquiries, contact the Met Office (https://www.metoffice.gov.uk/about-us/contact). Finally, terrain elevation data for the terrain maps are available from the Consultative Group on International Agricultural Research Consortium for Spatial Information (CGIAR-CSI; https://srtm.csi.cgiar.org/srtmdata/) and country shapefiles are from the Centers for Disease Control and Prevention (CDC; https://www.cdc.gov/epiinfo/support/downloads/shapefiles.html).
Footnotes
Monthly IMERG and E-OBS precipitation patterns are also analyzed (not shown), with similar results.
For clarity, convective precipitation accumulations over the Pyrenees, Alps, and Carpathians in IMERG are of similar magnitude to accumulations over the Adriatic, Mediterranean, North, and Baltic Seas in IMERG. Recall the different shading ranges in Figs. 12 and 15.
Note a different map projection is used due to a Python projection shading bug.
REFERENCES
Alhammoud, B., C. Claud, B. M. Funatsu, K. Béranger, and J.-P. Chaboureau, 2014: Patterns of precipitation and convection occurrence over the Mediterranean Basin derived from a decade of microwave satellite observations. Atmosphere, 5, 370–398, https://doi.org/10.3390/atmos5020370.
Allen, J. T., 2018: Climate Change and Severe Thunderstorms. Oxford University Press, 70 pp, https://doi.org/10.1093/acrefore/9780190228620.013.62.
Anagnostou, E. N., T. Chronis, and D. P. Lalas, 2002: New receiver network advances long-range lightning monitoring. Eos Trans., 83, 589–595, https://doi.org/10.1029/2002EO000403.
Anderson, G., and D. Klugmann, 2014: A European lightning density analysis using 5 years of ATDnet data. Nat. Hazards Earth Syst. Sci., 14, 815–829, https://doi.org/10.5194/nhess-14-815-2014.
Azorin-Molina, C., and D. Chen, 2009: A climatological study of the influence of synoptic-scale flows on sea breeze evolution in the Bay of Alicante (Spain). Theor. Appl. Climatol., 96, 249–260, https://doi.org/10.1007/s00704-008-0028-2.
Bandhauer, M., and Coauthors, 2022: Evaluation of daily precipitation analyses in E-OBS (v19.0e) and ERA5 by comparison to regional high-resolution datasets in European regions. Int. J. Climatol., 42, 727–747, https://doi.org/10.1002/joc.7269.
Battan, L. J., and J. B. Theiss, 1973: Wind gradients and variance of Doppler spectra in showers viewed horizontally. J. Appl. Meteor. Climatol., 12, 688–693, https://doi.org/10.1175/1520-0450(1973)012<0688:WGAVOD>2.0.CO;2.
Bech, J., J. Arús, S. Castán, N. Pineda, T. Rigo, J. Montanyà, and O. van der Velde, 2015: A study of the 21 March 2021 tornadic quasi linear convective system on Catalonia. Atmos. Res., 158–159, 192–209, https://doi.org/10.1016/j.atmosres.2014.08.009.
Bechini, R., D. Giaiotti, A. Manzato, F. Stel, and S. Micheletti, 2001: The June 4th 1999 severe weather episode in San Quirino, Italy: A tornado event? Atmos. Res., 56, 213–232, https://doi.org/10.1016/S0169-8095(00)00074-0.
Becker, A., P. Finger, A. Meyer-Christoffer, B. Rudolf, K. Schamm, U. Schneider, and M. Ziese, 2013: A description of the global land-surface precipitation data products of the Global Precipitation Climatology Centre with sample applications including centennial (trend) analysis from 1901-present. Earth Syst. Sci. Data, 5, 71–99, https://doi.org/10.5194/essd-5-71-2013.
Bertato, M., D. B. Giaiotti, A. Manzato, and F. Stel, 2003: An interesting case of tornado in Friuli-Northeastern Italy. Atmos. Res., 67–68, 3–21, https://doi.org/10.1016/S0169-8095(03)00043-7.
Brooks, H. E., 2013: Severe thunderstorms and climate change. Atmos. Res., 123, 129–138, https://doi.org/10.1016/j.atmosres.2012.04.002.
Buontempo, C., and Coauthors, 2020: Fostering the development of climate services through Copernicus Climate Change Service (C3S) for agriculture applications. Wea. Climate Extremes, 27, 100226, https://doi.org/10.1016/j.wace.2019.100226.
Buzzi, A., and P. P. Alberoni, 1992: Analysis and numerical modelling of a frontal passage associated with thunderstorm development over the Po Valley and the Adriatic Sea. Meteor. Atmos. Phys., 48, 205–224, https://doi.org/10.1007/BF01029569.
Cacciamani, C., F. Battaglia, P. Patruno, L. Pomi, A. Selvini, and S. Tibaldi, 1995: A climatological study of thunderstorm activity in the Po Valley. Theor. Appl. Climatol., 50, 185–203, https://doi.org/10.1007/BF00866116.
Campos, R. M., C. B. Gramcianinov, R. de Camargo, and P. L. da Silva Dias, 2022: Assessment and calibration of ERA5 severe winds in the Atlantic Ocean using satellite data. Remote Sens., 14, 4918, https://doi.org/10.3390/rs14194918.
Collino, E., P. Bonelli, and L. Gilli, 2009: ST-AR (STorm-Archive): A project developed to assess the ground effects of severe convective storms in the Po Valley. Atmos. Res., 93, 483–489, https://doi.org/10.1016/j.atmosres.2008.10.021.
Comin, A. N., M. M. Miglietta, U. Rizza, O. C. Acevedo, and G. A. Degrazia, 2015: Investigation of sea-breeze convergence in Salento Peninsula (southeastern Italy). Atmos. Res., 160, 68–79, https://doi.org/10.1016/j.atmosres.2015.03.010.
Cornes, R. C., G. van den Schrier, E. J. M. van den Besselaar, and P. D. Jones, 2018: An ensemble version of the E-OBS temperature and precipitation data sets. J. Geophys. Res. Atmos., 123, 9391–9409, https://doi.org/10.1029/2017JD028200.
Da Silva, N. A., and J. O. Haerter, 2023: The precipitation characteristics of mesoscale convective systems over Europe. J. Geophys. Res. Atmos., 128, e2023JD039045, https://doi.org/10.1029/2023JD039045.
Davolio, S., A. Volonté, A. Manzato, A. Pucillo, A. Cicogna, and M. E. Ferrario, 2016: Mechanisms producing different precipitation patterns over north-eastern Italy: Insights from HyMeX-SOP1 and previous events. Quart. J. Roy. Meteor. Soc., 142, 188–205, https://doi.org/10.1002/qj.2731.
del Moral, A., M. del Carmen Llasat, and T. Rigo, 2017: Identification of anomalous motion of thunderstorms using daily rainfall fields. Atmos. Res., 185, 92–100, https://doi.org/10.1016/j.atmosres.2016.11.001.
del Moral, A., M. de Carmen Llasat, and T. Rigo, 2020: Connecting flash flood events with radar-derived convective storm characteristics on the northwestern Mediterranean coast: Knowing the present for better future scenarios adaptation. Atmos. Res., 238, 104863, https://doi.org/10.1016/j.atmosres.2020.104863.
Enno, S.-E., G. Anderson, and J. Sugier, 2016: ATDnet detection efficiency and cloud lightning detection characteristics from comparison with the HyLMA during HyMeX SOP1. J. Atmos. Oceanic. Technol., 33, 1899–1911, https://doi.org/10.1175/JTECH-D-15-0256.1.
Enno, S.-E., J. Sugier, R. Alber, and M. Seltzer, 2020: Lightning flash density in Europe based on 10 years of ATDnet data. Atmos. Res., 235, 104769, https://doi.org/10.1016/j.atmosres.2019.104769.
Fairman, J. G., Jr., D. M. Schultz, D. J. Kirshbaum, S. L. Gray, and A. I. Barrett, 2017: Climatology of size, shape, and intensity of precipitation features over Great Britain and Ireland. J. Hydrometeor., 18, 1595–1615, https://doi.org/10.1175/JHM-D-16-0222.1.
Feudale, L., and A. Manzato, 2014: Cloud-to-ground lightning distribution and its relationship with orography and anthropogenic emissions in the Po Valley. J. Appl. Meteor. Climatol., 53, 2651–2670, https://doi.org/10.1175/JAMC-D-14-0037.1.
Fibbi, L., M. Chiesi, M. Moriondo, M. Bindi, G. Chirici, D. Papale, B. Gozzini, and F. Maselli, 2016: Correction of a 1 km daily rainfall dataset for modelling forest ecosystem processes in Italy. Meteor. Appl., 23, 294–303, https://doi.org/10.1002/met.1554.
Frei, C., and C. Schär, 1998: A precipitation climatology of the Alps from high-resolution rain-gauge observations. Int. J. Climatol., 18, 873–900, https://doi.org/10.1002/(SICI)1097-0088(19980630)18:8<873::AID-JOC255>3.0.CO;2-9.
Frei, C., and F. A. Isotta, 2019: Ensemble spatial precipitation analysis from rain guage data: Methodology and application in the European Alps. J. Geophys. Res. Atmos., 124, 5757–5778, https://doi.org/10.1029/2018JD030004.
Gallus, W. A., Jr., N. A. Snook, and E. V. Johnson, 2008: Spring and summer severe weather reports over the Midwest as a function of convective mode: A preliminary study. Wea. Forecasting, 23, 101–113, https://doi.org/10.1175/2007WAF2006120.1.
Godart, A., S. Anquetin, E. Leblois, and J.-D. Creutin, 2011: The contribution of orographically driven banded precipitation to the rainfall climatology of a Mediterranean region. J. Appl. Meteor. Climatol., 50, 2235–2246, https://doi.org/10.1175/JAMC-D-10-05016.1.
González-Alemán, J. J., D. Insua-Costa, E. Bazile, S. González-Herrero, M. M. Miglietta, P. Groenemeijer, and M. G. Donat, 2023: Anthropogenic warming had a crucial role in triggering the historic and destructive Mediterranean derecho in summer 2022. Bull. Amer. Meteor. Soc., 104, E1526–E1532, https://doi.org/10.1175/BAMS-D-23-0119.1.
Güttler, I., I. Stepanov, Č. Branković, G. Nikulin, and C. Jones, 2015: Impact of horizontal resolution on precipitation in complex orography simulated by the regional climate model RCA3. Mon. Wea. Rev., 143, 3610–3627, https://doi.org/10.1175/MWR-D-14-00302.1.
Haylock, M. R., N. Hofstra, A. M. G. Klein Tank, E. J. Klok, P. D. Jones, and M. New, 2008: A European daily high-resolution gridded data set of surface temperature and precipitation for 1950–2006. J. Geophys. Res., 113, D20119, https://doi.org/10.1029/2008JD010201.
Holzer, A. M., T. M. E. Schreiner, and T. Púčik, 2018: A forensic reanalysis of one of the deadliest European tornadoes. Nat. Hazards Earth Syst. Sci., 18, 1555–1565, https://doi.org/10.5194/nhess-18-1555-2018.
Hou, A. Y., and Coauthors, 2014: The Global Precipitation Measurement Mission. Bull. Amer. Meteor. Soc., 95, 701–722, https://doi.org/10.1175/BAMS-D-13-00164.1.
Huffman, G. J., D. T. Bolvin, E. J. Nelkin, and J. Tan, 2020a: Integrated Multi-satellitE Retrievals for GPM (IMERG) technical documentation. NASA/GSFC Tech. Rep., 97 pp., https://gpm.nasa.gov/sites/default/files/2023-07/IMERG_TechnicalDocumentation_final_230713.pdf.
Huffman, G. J., and Coauthors, 2020b: NASA Global Precipitation Measurement (GPM) Integrated Multi-satellitE Retrievals for GPM (IMERG) Algorithm Theoretical Basis Doc. (ATBD), version 6, 39 pp., https://gpm.nasa.gov/sites/default/files/2020-05/IMERG_ATBD_V06.3.pdf.
Huuskonen, A., E. Saltikoff, and I. Holleman, 2014: The operational weather radar network in Europe. Bull. Amer. Meteor. Soc., 95, 897–907, https://doi.org/10.1175/BAMS-D-12-00216.1.
Illingworth, A., and R. Thompson, 2022: Reconciling the differences between radar and rain gauges, both are correct, but radar is usually more representative. ERAD Conf., Locarno, Switzerland, Federal Office of Meteorology and Climatology MeteoSwiss and L’École Polytechnique Fédérale de Lausanne, 118, https://www.erad2022.ch/_files/ugd/25a7b1_23b223c508be440ca402c5a29bba166e.pdf.
Isotta, F. A., and Coauthors, 2014: The climate of daily precipitation in the Alps: Development and analysis of a high-resolution grid dataset from pan-Alpine rain-gauge data. Int. J. Climatol., 34, 1657–1675, https://doi.org/10.1002/joc.3794.
Jansa, A., and Coauthors, 2014: MEDEX: A general overview. Nat. Hazards Earth Syst. Sci., 14, 1965–1984, https://doi.org/10.5194/nhess-14-1965-2014.
Katona, B., P. Markowski, C. Alexander, and S. Benjamin, 2016: The influence of topography on convective storm environments in the eastern United States as deduced from the HRRR. Wea. Forecasting, 31, 1481–1490, https://doi.org/10.1175/WAF-D-16-0038.1.
Klok, E. J., and A. M. G. Klein Tank, 2009: Updated and extended European dataset of daily climate observations. Int. J. Climatol., 29, 1182–1191, https://doi.org/10.1002/joc.1779.
Kotroni, V., and K. Lagouvardos, 2016: Lightning in the Mediterranean and its relation with sea-surface temperature. Environ. Res. Lett., 11, 034006, https://doi.org/10.1088/1748-9326/11/3/034006.
Kummerow, C., and Coauthors, 2000: The status of the Tropical Rainfall Measuring Mission (TRMM) after two years in orbit. J. Appl. Meteor., 39, 1965–1982, https://doi.org/10.1175/1520-0450(2001)040<1965:TSOTTR>2.0.CO;2.
Kunz, M., U. Blahak, J. Handwerker, M. Schmidberger, H. Jürgen Punge, S. Mohr, E. Fluck, and K. M. Bedka, 2018: The severe hailstorm in southwest Germany on 28 July 2013: Characteristics, impacts, and meteorological conditions. Quart. J. Roy. Meteor. Soc., 144, 231–250, https://doi.org/10.1002/qj.3197.
Llasat, M. C., and M. Puigcerver, 1997: Total rainfall and convective rainfall in Catalonia, Spain. Int. J. Climatol., 17, 1683–1695, https://doi.org/10.1002/(SICI)1097-0088(199712)17:15<1683::AID-JOC220>3.0.CO;2-Q.
Llasat, M. C., R. Marcos, M. Turco, J. Gilabert, and M. Llasat-Botija, 2016: Trends in flash flood events versus convective precipitation in the Mediterranean region: The case of Catalonia. J. Hydrol., 541, 24–37, https://doi.org/10.1016/j.jhydrol.2016.05.040.
Llasat, M. C., M. del Moral, M. Cortès, and T. Rigo, 2021: Convective precipitation trends in the Spanish Mediterranean region. Atmos. Res., 257, 105581, https://doi.org/10.1016/j.atmosres.2021.105581.
Lombardo, K. A., and B. A. Colle, 2010: The spatial and temporal distribution of organized convective structures over the northeast and their ambient conditions. Mon. Wea. Rev., 138, 4456–4474, https://doi.org/10.1175/2010MWR3463.1.
Lopez, P., 2014: Comparison of ODYSSEY precipitation composites to SYNOP rain gauges and ECMWF model. ECMWF Tech. Memo. 717, 17 pp.
Manzato, A., S. Serafin, M. M. Miglietta, D. Krishbaum, and W. Schultz, 2022: A pan-alpine climatology of lightning and convection initiation. Mon. Wea. Rev., 150, 2213–2230, https://doi.org/10.1175/MWR-D-21-0149.1.
Mehta, A. V., and S. Yang, 2008: Precipitation climatology over Mediterranean Basin from ten years of TRMM measurements. Adv. Geosci., 17, 87–91, https://doi.org/10.5194/adgeo-17-87-2008.
Mohr, S., J. Wandell, S. Lenggenhager, and O. Martius, 2019: Relationship between atmospheric blocking and warm-season thunderstorms over western and central Europe. Quart. J. Roy. Meteor. Soc., 145, 3040–3056, https://doi.org/10.1002/qj.3603.
Mohr, S., and Coauthors, 2020: The role of large-scale dynamics in an exceptional sequence of severe thunderstorms in Europe May–June 2018. Wea. Climate Dyn., 1, 325–348, https://doi.org/10.5194/wcd-1-325-2020.
Navarro, A., E. García-Ortega, A. Merino, J. L. Sánchez, C. Kummerow, and F. J. Tapiador, 2019: Assessment of IMERG precipitation estimates over Europe. Remote Sens., 11, 2470, https://doi.org/10.3390/rs11212470.
Overeem, A., E. van den Besselaar, G. van der Schrier, J. Meirink, E. van der Plas, and H. Leijnse, 2022: EURADCLIM: The European climatological gauge-adjusted radar precipitation dataset (1-h accumulations). KNMI Data Platform, accessed 7 November 2023, https://doi.org/10.21944/7ypj-wn68.
Overeem, A., E. van den Besselaar, G. van der Schrier, J. F. Meirink, E. van der Plas, and H. Leijnse, 2023: EURADCLIM: The European climatological high-resolution gauge-adjusted radar precipitation dataset. Earth Syst. Sci. Data, 15, 1441–1464, https://doi.org/10.5194/essd-15-1441-2023.
Paccagnella, T., S. Tibaldi, R. Buizza, and S. Scoccianti, 1992: High-resolution numerical modelling of convective precipitation over northern Italy. Meteor. Atmos. Phys., 50, 143–163, https://doi.org/10.1007/BF01025510.
Park, S., M. Berenguer, and D. Sempere-Torres, 2019: Long-term analysis of gauge-adjusted radar rainfall accumulations at European scale. J. Hydrol., 573, 768–777, https://doi.org/10.1016/j.jhydrol.2019.03.093.
Parker, M. D., and J. C. Knievel, 2005: Do meteorologists suppress thunderstorms?: Radar-derived statistics and the behavior of moist convection. Bull. Amer. Meteor. Soc., 86, 341–358, https://doi.org/10.1175/BAMS-86-3-341.
Pavan, V., and Coauthors, 2019: High resolution climate precipitation analysis for north-central Italy, 1961–2015. Climate Dyn., 52, 3435–3453, https://doi.org/10.1007/s00382-018-4337-6.
Peura, M., M. Martet, A. Huuskonen, L. Delobbe, B. Lipovscak, H. Leijnse, and E. Saltikoff, 2017: OPERA - Harmonizing the European weather radar network. 38th Conf. on Radar Meteorology, Chicago, IL, Amer. Meteor. Soc., 168, https://ams.confex.com/ams/38RADAR/webprogram/Manuscript/Paper320437/AMS_OPERA_Peura_etal_2017_extended.pdf.
Prakash, S., A. K. Mitra, A. AghaKouchak, Z. Liu, H. Norouzi, and D. S. Pai, 2018: A preliminary assessment of GPM-based multi-satellite precipitation estimates over a monsoon dominated region. J. Hydrol., 556, 865–876, https://doi.org/10.1016/j.jhydrol.2016.01.029.
Pucillo, A., M. M. Miglietta, K. Lombardo, and A. Manzato, 2020: Application of a simple analytical model to severe winds produced by a bow echo like storm in Northeast Italy. Meteor. Appl., 27, e1868, https://doi.org/10.1002/met.1868.
Rakov, V. A., and M. A. Uman, 2003: Lightning: Physics and Effects. 1st ed. Cambridge University Press, 687 pp.
Ramsauer, T., T. Weiß, and P. Marzahn, 2018: Comparison of the GPM IMERG final precipitation product to RADOLAN weather radar data over the topographically and climatically diverse Germany. Remote Sens., 10, 2029, https://doi.org/10.3390/rs10122029.
Rashid, S. A., M. J. Iqbal, and M. A. Hussain, 2012: Impact of north-south shift of Azores high on summer precipitation over North West Europe. Int. J. Geosci., 3, 992–999, https://doi.org/10.4236/ijg.2012.325099.
Rebora, N., and Coauthors, 2013: Extreme rainfall in the Mediterranean: What can we learn from observations? J. Hydrometeor., 14, 906–922, https://doi.org/10.1175/JHM-D-12-083.1.
Reeves, H. D., and Y.-L. Lin, 2006: Effect of stable layer formation over the Po Valley on the development of convection during MAP IOP-8. J. Atmos. Sci., 63, 2567–2584, https://doi.org/10.1175/JAS3759.1.
Saltikoff, E., 2015: Survey on implementation of INSPIRE directive among OPERA members. OPERA Internal Rep. 7 pp.
Saltikoff, E., M. Kurri, H. Leijnse, S. Barbosa, and K. Stiansen, 2017: Maintenance keeps radars running. Bull. Amer. Meteor. Soc., 98, 1833–1840, https://doi.org/10.1175/BAMS-D-16-0095.1.
Saltikoff, E., and Coauthors, 2019: OPERA the radar project. Atmosphere, 10, 320, https://doi.org/10.3390/atmos10060320.
Sanders, R. A., 1953: Blocking highs over the eastern North Atlantic Ocean and western Europe. Mon. Wea. Rev., 81, 67–73, https://doi.org/10.1175/1520-0493(1953)081<0067:BHOTEN>2.0.CO;2.
Tapiador, F. J., A. Navarro, E. García-Ortega, A. Merino, J. L. Sánchez, C. Marcos, and C. Kummerow, 2020: The contribution of rain gauges in the calibration of the IMERG product: Results from the first validation over Spain. J. Hydrometeor., 21, 161–182, https://doi.org/10.1175/JHM-D-19-0116.1.
Taszarek, M., and Coauthors, 2019a: A climatology of thunderstorms across Europe from a synthesis of multiple data sources. J. Climate, 32, 1813–1837, https://doi.org/10.1175/JCLI-D-18-0372.1.
Taszarek, M., and Coauthors, 2019b: Derecho evolving from a mesocyclone—A study of 11 August 2017 severe weather outbreak in Poland: Event analysis and high-resolution simulation. Mon. Wea. Rev., 147, 2283–2306, https://doi.org/10.1175/MWR-D-18-0330.1.
Taszarek, M., J. T. Allen, P. Groenemeijer, R. Edwards, H. E. Brooks, V. Chmielewski, and S.-E. Enno, 2020a: Severe convective storms across Europe and the United States. Part I: Climatology of lightning, large hail, severe wind, and tornadoes. J. Climate, 33, 10 239–10 261, https://doi.org/10.1175/JCLI-D-20-0345.1.
Taszarek, M., J. T. Allen, T. Púčik, K. A. Hoogewind, and H. E. Brooks, 2020b: Severe convective storms across Europe and the United States. Part II: ERA5 environments associated with lightning, large hail, severe wind, and tornadoes. J. Climate, 33, 10 263–10 286, https://doi.org/10.1175/JCLI-D-20-0346.1.
Tokay, A., and D. A. Short, 1996: Evidence from tropical raindrop spectra of the origin of rain from stratiform versus convective clouds. J. Appl. Meteor. Climatol., 35, 355–371, https://doi.org/10.1175/1520-0450(1996)035<0355:EFTRSO>2.0.CO;2.
Tramblay, Y., H. Feki, P. Quintana-Seguí, and J. A. Guijarro, 2019: The SAFRAN daily gridded precipitation product in Tunisia (1979–2015). Int. J. Climatol., 39, 5830–5838, https://doi.org/10.1002/joc.6181.
Trapero, L., J. Bech, and J. Lorente, 2013: Numerical modelling of heavy precipitation events over eastern Pyrenees: Analysis of orographic effects. Atmos. Res., 123, 368–383, https://doi.org/10.1016/j.atmosres.2012.09.014.
van Delden, A., 2001: The synoptic setting of thunderstorms in western Europe. Atmos. Res., 56, 89–110, https://doi.org/10.1016/S0169-8095(00)00092-2.
Wilhelm, J., S. Mohr, H. Jürgen Punge, B. Mühr, M. Schmidberger, J. E. Daniell, K. M. Bedka, and M. Kunz, 2020: Severe thunderstorms with large hail across Germany in June 2019. Weather, 76, 228–237, https://doi.org/10.1002/wea.3886.
