How changements in the circulation patterns specific to the solid precipitations can affect these meteorological events in the Alpine stations of the Mediterranean region ? Use of the ERA5 reanalyses

doi:10.21203/rs.3.rs-3008484/v1

Download PDF

Research Article

How changements in the circulation patterns specific to the solid precipitations can affect these meteorological events in the Alpine stations of the Mediterranean region ? Use of the ERA5 reanalyses

https://doi.org/10.21203/rs.3.rs-3008484/v1

This work is licensed under a CC BY 4.0 License

Journal Publication

published 24 Nov, 2023

Read the published version in Theoretical and Applied Climatology →

You are reading this latest preprint version

A classification was achieved to obtain some daily atmospheric circulation types (days that have a similar atmospheric circulation pattern), capable of reproducing sequences specific to the solid precipitations’ events that occur in the mountains of the southern Alps (Mediterranean region). These atmospheric circulation types were statistically studied to get the evolution of the frequency of the synoptic conditions favorable and unfavorable to solid precipitations. The classification produced six atmospheric circulation types, including two synoptic snow classes and two dry classes. The South circulation with a trough on western Europe (STW) is the major snow classes. STW reproduced 62% of the total solid precipitations’ events, 68% of those with a medium intensity level and 100% of those with the high intensity level, which shows the effectiveness of the approach. The total number of days of presence per decades of STW, decreased between 1950 and 2019. This decreasing frequency of occurrence also applies to the other synoptic class of snow, the Depression over Central Europe (DCE). Conversely, the West anticyclonic (WA) and Southwest anticyclonic (SWA), did not produce any solid precipitations events and indicated the opposite logic, with an increase in their occurrences for the same period. This could lead to a strong change in the climate of the South of France, with major impact on the thickness and duration of snow cover. A sharp decrease in snow cover would have heavy economic consequences, the water cycle would be modified, the management of water resources could then be severely called into question and more worryingly, it would increase the intensity of periods of drought. The most plausible explanation for this phenomenon would come from the spread of Hadley's cell, which would push the Azores anticyclone further north and leads to a higher frequency of occurrence of WA and SWA.

Solid precipitation

Alpine stations

Circulation patterns

ERA5 reanalyses

Classification

Snow is an important part of the environmental system and generates economic attractiveness (Stoffel & Corona, 2018; Berard-Chenu et al., 2022). The duration, distribution, and spatial extent of the natural snow cover is essential for the hydrological cycle. It is also determinant for vegetation or agriculture, and vital for the communities settled in the mountains through the economy of the winter tourism (Barbier, 1982; Bokhorst et al., 2016; François et al., 2016; Croce et al., 2018; Joye, 2019; Zhang et al., 2019; Franco, 2021; Barry & Gan., 2022).

This paper is interested in the solid precipitations (SP) of the South of France, those of the Alpine stations located in the French department of the Alpes-Maritimes (Southern Alps). In this Mediterranean region, contemporary climate change threatens thickness and extent of the snow cover, with three significant issues. Firstly, from an economic point of view, snow generates through winter sports tourism, 20 million euros turnover per season and 15 000 jobs for the Alpes-Maritimes (Côte d'Azur tourism observatory statistical - study from 2010 to 2015). The natural snow cover, which results from SP, is one of the determinants of this economic activity, although, snow at ski resorts is now well managed with modern techniques of grooming and production (Morin et al., 2022). However, an important condition is to have enough precipitation to fill the retention lakes and subsequently supply the snow cannons, which in turn require sufficiently low temperatures to operate. The other essential condition is to have intense SP at the beginning of the season to manage the slopes, thanks to the damage (Hanzer et al., 2020). The second issue is the management of water resources in the valleys. A reduction of the natural snow cover and an acceleration in the snowmelt in spring, alter the flow of rivers and their seasonality (Lastrada et al., 2021). There are therefore some impacts not only on ecosystems, but also on water management for farmers, dam managers, hydroelectricity producers and potentially on drinking water (Marnezy A., 2008). The third issue is an aggravation of the intensity of drought periods, since the natural snow cover acts as a buffer reservoir, by storing water in solid form during the winter and redistributing it in liquid form in the environmental system during the summer (Beniston, 2012). Moreover, periods of drought are increasingly marked in the region in the last years, for instance, the Alpes-Maritimes faced an exceptional situation in the summer 2022, leading to strong restrictions on the use of water for the inhabitants, but also to ecological disasters with several lakes and rivers dried up.

Evidence of contemporary climate change is clear, it affects all regions of the world by altering natural processes, modifying precipitation patterns, melting glaciers, rising sea levels, etc. The SP is a meteorological event formed when the atmosphere has a temperature below or close to 0°C (Klein, 2018). The Southern Alps, like the rest of the planet, have experienced a rise in the air temperature, since pre-industrial times (part 1 of the 6th IPCC report, 2021). There should be a continuation of warming at work for the middle and end of the 21st century, for scenarios with low Greenhouse Gas (GHG) emissions and an accentuation of warming for scenarios with high GHG emissions, with a real impact on the rain-snow limit pushed back higher in altitude. Climate change does not only impact the rain-snow limit, but it could also lead to a change in the atmospheric circulation, which in turn would conduct to changes in the occurrence of precipitation patterns (part 3 of the 6th IPCC report, 2021). In this sense, the atmospheric circulation, by redistributing at the planetary level the energy coming from the sun and the energy surplus coming from the intensification of the greenhouse effect is described as the engine of the weather for the past, present, and future period (Pédelaborde, 1956; Peña and Schulte, 2019). Then, the SP is the result of the action of an Atmospheric Circulation Type (ACT), which acts on the local weather in synergy with the substratum, the topography, etc. (Planchon et al., 2015). Some recent studies have focused on the relationship between the SP and the synoptic condition. Esteban et al., (2005) determined the synoptic situations responsible for heavy SP in Andorra in the Pyrenees. The authors used a combination of a Principal Component Analysis (PCA) and a K-means clustering to obtain a synoptic classification of days with heavy SP. Capozzi et al., (2022) in the Italian Apennines, Irannezhad et al., (2017) in Finland or even Merino et al., (2014) in Spain, carried out similar works.

The main objective of this study is to compute a classification of the atmospheric circulation specific to SP, especially the intense SP. The classification method must be reproducible to other geographical areas and the originality lies in the use of several atmospheric variables in its composition, which has so far never been done. The secondary objective is to statistically analyze the occurrence of synoptic conditions specific to the SP, to observe how the nival conditions in mid-latitude European mountains evolved. This work deals with the relationships between large-scale atmospheric flow and local SP, in the context of contemporary climate change. This is an additional tool to the current climate information available, to guide decision-making processes. Indeed, it is recognized that the outputs of the projection models include biases in the representation of certain sensitive weather variables, such as precipitation, but are relatively robust on typical circulation variables (Boé and Terray, 2008, Barrier et al., 2014; Pathak et al., 2019; Lavers et al., 2021).

2.1 Study area and solid precipitation observations

The Weather Stations Network (WSN) contains five stations: Isola 2000 (Y: 44.18, X: 7.15; Z: 1910 m), Valberg (Y: 44.09, X: 6.92; Z: 1630 m), Auron (Y: 44.22, X: 6.93; Z: 1910 m), Val Casterino (Y: 44.09, X: 7.50; Z: 1554 m) and Estenc (Y: 44.23, X: 6.75; Z: 1850 m). They are positioned on a NW/SE axis, in the North of the French department of the Alpes-Maritimes and have an attitudinal diversity (cf. Figure 1). The site enjoys a particular topographical context since it is at a geodesic distance of 40 km from the Mediterranean Sea. The influence of this reservoir of warm water throughout the year (the water temperature rarely drops below 12 ◦C even in winter) on the land borders, is an obvious fact, with as a consequence, the important total rainfall during Mediterranean episodes (Kyselý et al., 2012; Piras et al., 2016; Chochon et al., 2021; Carrega et al., 2021).

These five stations are the only ones in the department to have snow surveys covering periods post 2000s, this justifies their use. They measure in centimeters the amount of the fresh SP fallen in 24 hours. This data will be named in this paper “24-hour SP” and comes from the official service of meteorology and climatology in France (Météo-France). The measurements are done once a day at 6:00 UTC. The 24-hour SP is counted from 6:00 UTC Day-D to 6:00 UTC Day-D + 1. In other words, all the D-Day 24-hour SP are measured the next morning at 6:00 UTC. The measurements are made manually from a snowy table. It is a square wooden board on which is fixed a rod of 30 cm high. It is then a question of inserting a ruler graduated in centimeters to the board, to know the height of the 24-hour SP.

2.1.1 Data quality control of the 24-hour solid precipitations

The consistency and reliability of a time series of observations can be compromised by several errors and artifacts, related to human errors in data collection, inaccuracies in digitization procedures and gaps in the data series (Capozzi et al., 2022). However, in the meteorological and climatological fields, there is no unambiguous and universal quality control strategy, since, depending on the type and number of data, different methods can be applied (Ashcroft et al., 2018). In the case of this work, the data verification methods use the work of Capozzi et al., (2022) as a reference. The first control is to verify whether in all five series of observations (Estenc, Isola 2000, Valberg, Auron, Val Casterino) outliers appear. The 24-hour SP of the five stations are sorted in ascending order and the values above the 95th percentile are slightly studied. If one of these values is abnormally higher in a station, compared to neighboring stations that are geographically close without having a plausible explanation, it is then rejected. In the case of this report, no outlying values were detected. The second control aims to verify the homogeneity of the five series of observations to identify an artificial change or tendencies that could affect the results of the analysis. A break in the homogeneity of data can be attributed to several factors, including instrumentation error, changes in the environment adjacent to the instrument, and human mishandling (Capozzi et al., 2022). The test of Pettitt (Pettitt, 1979), then a comparison of SP trends between neighboring stations was applied to the five-observation series and no break in homogeneity linked to an artefact was detected.

2.1.2 Intensity classes of the 24-hour solid precipitations

The determination of the intensity of the 24-hour SP is achieved through a classification, to characterize the limits of each type (Martin, 2008) and obtain groups with minimized variance between individuals in the same group, and maximum variance between different groups (Creusier et al., 2014). One of the most widely used analyses by climatologists remains the combination of a Principal Component Analysis (PCA), which simplifies the information in the dataset, and an Hierarchical Clustering Analyze (HCA), which classifies statistical individuals into different groups (Chiadjeu et al., 2016).

This combination is used on the 24-hour SP. The common temporal extension of the 24-hour SP coming from the WSN extends from 1997 to 2019, or 23 years, composed of a seasonal winter-spring set grouping five months (DJFMA): December, January, February, March and April. The season of the SP in the South of France extends from November to April (Matiu et al., 2021), but, due to missing data for November, the study season has been restricted to five months (DJFMA). This period (1997–2019 DJFMA), include 3 598 days. However, only days with a 24-hour SP measured in all the five stations of the WSN are retained, reducing the period to 243 days. Thus, the first step is to perform a PCA over this matrix which includes 243 rows (days) and 5 columns (stations). Kaiser's rule (1961) is then applied to keep only the components whose eigenvalues score is greater than 1. Finally, the result of the PCA (coordinates of Individuals) is used in the HCA, to distinguish several classes of intensity of 24-hour SP.

2.2 Atmospheric variables

2.2.1 Atmospheric circulation types

The Solid Precipitation (SP) is dependent on climatic conditions (Ye and Rawlins., 2013; Smith et al., 2020) and atmospheric Variables (AV) such as geopotential height, atmospheric air temperature, specific humidity, or others, influencing them (Viceto et al., 2017, Guerin & Martin, 2022). Some AV act on the formation process, others on the transport mechanisms (Martin & Carrega., 2011, Guerin & Martin, 2022). The variations of the AV follow cyclical rhythms and weather conditions are sometimes repetitive or persistent (Martin, 2008). This makes it possible to group days that have a similar atmospheric circulation pattern, without necessarily being successive, in this study they are named ACT, for Atmospheric Circulation Types. The ACT generate the type of weather (sunny, rainy, warm, cold etc.), or rather is located upstream (Carrega, 2004). The main objective of this paper is to determine the ACT which causes intense SP in the study area, to do this it is necessary to use large scale climate data.

2.2.2 The ERA5 reanalysis

The AV that will be used for the classification and production of the ACT are taken from the reanalyses ERA5 provided by the European Centre for Medium-Range Weather Forecasts (ECMWF). They are at the interface between observation and modeling (Parker, 2016; Zuo et al., 2019; Nogueira, 2020) and due to the integration of measured data, the reanalysis seems closer to the observed climatology than to the climatology of the model (Poccard, 2000). Exploitable since 1950, the ERA5 reanalysis has a horizontal resolution of 0.25 ° x 0.25° and an hourly temporal time step (Silber et al., 2019). They have been chosen because they have a high resolution for the WSN (the highest at the start of the work), produce a complete hourly data series for a wide choice of AV at different levels of pressure and have a geographical window large enough to visualize the depressions of high latitudes and the anticyclones of low latitudes. Also, the data assimilation method with the 4D-VAR technology, shows more robust trends in climate variables (Mahto et al., 2019).

2.2.3 Selection of the atmospherics variables

The AV, available online, provide an abundant source of data. The major obstacle is not to get lost in this plethora. Indeed, to build a robust classification, it is necessary to make a choice, depending on the intensity of the statistical relationships. The method of selection of the AV involves carrying out Bravais-Pearson correlation tests (r) between the 24-hour SP and 10 AV, consisting for the most part of three pressure levels, which gives a total of 72 combinations (cf. Table 1). The ERA5 can be downloaded hourly, while the measurements of the 24-hour SP from the weather station are taken at 6:00 UTC the next day. The choice of the hourly steps is therefore essential to increase the chances of not missing the synoptic conditions specific to the snow event. For the study, it was decided to test four-time steps: 6:00 UTC, 12:00 UTC, 18:00 UTC and 00:00 UTC.

Thus, for each reanalysis pixel including a WSN station, the values of the AV are extracted, then r tests are produced between the 72 combinations and the 24-hour observation data. Moreover, to verify that the ERA5 pixels correctly represent the physical environment around the station, correlations were also carried out on the other pixels neighboring the station. The period common to these two sets of variables includes 2380 days (1997–2019 DJFMA) which serve as the basis for this work. Following the calculation of r, two controls must be carried out. It must be checked that the two sets of data do not differ from a Gaussian function, and it is necessary to ensure that the value of r is outside the rates compatible with chance (Pédelaborde, 1956). It is interesting to rely on the work of the German meteorologist Baur (1930) to support the significance of the r tests. According to his work, if the data provides more than 620 values in each series, the r must exceed 0.122 in absolute value, to reveal a statistical relationship.

Table 1

Atmospheric variables tested (ECMWF, ERA5).
Variables	Levels (hPa)	Time steps (Hour - UTC)	Combinations
Geopotential	500; 700; 850	00; 06; 12; 18	12
Specific humidity	500; 700; 850	00; 06; 12; 18	12
U-Component-Wind	500; 700; 850	00; 06; 12; 18	12
V-Component-Wind	500; 700; 850	00; 06; 12; 18	12
Air temperature	500; 700; 850	00; 06; 12; 18	12
Vertical velocity	500; 700; 850	00; 06; 12; 18	12
2 meters air temperature	2 meters	00; 06; 12; 18	4
Mean sea level pressure	sea level	00; 06; 12; 18	4
Surface pressure	surface	00; 06; 12; 18	4
Sea surface temperature	surface	00; 06; 12; 18	4

2.2.4 Window extraction of the atmospherics variables fields

Once the selection of the AV has been made, the data must be extracted. The AV fields are available at a latitudinal and longitudinal resolution of 0.25°. The extraction window used, extends from 30° South to 60° North of latitude and from 40° West to 30° East of longitude. This choice is motivated to leave the WSN as close as possible to the center of the space, while being slightly offset towards the East: the general circulation being oriented from West to East at this latitude, it is essential to consider reanalyzed atmospheric fields over the near Atlantic. Even, if they do not directly influence the SP of the WSN, they are frequently located upstream of the atmospheric circulation that concerns it. A total of 1 785 extraction points make up the extraction window of the AV. A maximum resolution (0.25°) is kept in the center of the map, while a lower resolution is used elsewhere (1.25°), to weight the grids, close to the South-East of France more heavily. Many windows have been tested with different resolutions and the Fig. 2 presents the window that provided the best result (smallest deviations in the patterns of circulation linked to the classification of AV fields, in comparison of the patterns linked to the SP classes). Each AV is extracted for the 1 785 grid points and for all day of DJFMA. All these days will be classified, according to the method described below.

2.2.5 Method of classification of atmospherics variables fields

The next step is to produce a classification of the daily AV fields to obtain some ACT with a combination of a PCA and an HCA. The methodology adopted in this work used the work of Martin (2008) as a reference. The first step is always to perform a PCA on the data matrices where the variables are the 1 785 grid points of the European window, and the individuals are the days. Several PCA are conducted, one for each AV previously selected. In each case, the Principal Components (PC) are retained according to the Kaiser’s rule: retain PC whose eigenvalues are greater than 1. The Kaiser rule is the most used approach to selecting the number of components and an eigenvalue of 1 means that the PC contains the same amount of information as a single variable (Nunnally and Bernstein, 1994). This technique has shown the smallest deviations in the patterns of circulation linked to the classification of AV fields, in comparison of the patterns linked to the SP classes. Then, all the main components selected from each PCA are brought together in a new matrix subject to an HCA, with an algorithm of Ward. The HCA will then gather the individuals iteratively to produce a dendrogram, and the inter-group inertia is used to partition the classification. Finally, two classifications are made. A first from 1997 to 2019, which corresponds to the control classification. Then a second from 1950 to 2019, based on the same method as the first, but which will ensure statistical monitoring of ACT over a period of 70 years. The Fig. 3 below, summarizes the different steps of the method used.

3.1 Classification of the solid precipitations

A combination of a PCA and an HCA is conducted over the 243 days when a 24-hour SP is present in the five stations of the WSN, to obtain a classification of the SP observations. The objective is to discretize the 24-hours in intensity class. The first four dimensions of the PCA were retained. They express 98% of the total inertia (cf. Figure 4) and a final inertia jump is observed between a cut in 3 or 4 classes, this justifies the choice made. The results of the PCA (the first four dimensions) are used to make an HCA, which shows three homogeneous and significant classes of SP, as illustrated in Fig. 5.

The mean of the amount of SP fell in 24 hours and the occurrence for each class are presented in Table 2. Class 1 (SP1) is composed of 24-hour SP sharing low values for all stations, with a mean of 2.3 cm fell in 24 hours and an occurrence of 129 days. Class 2 (SP2) has a mean value around 16.9 cm and an occurrence of 79 days, finally, class 3 (SP3) corresponds to high values, with 38 cm and an occurrence of 35 days.

Table 2

Mean amount of SP in cm per classes and occurrence 1997–2019 DJFMA (Météo-France).
Classes	Val Casterino	Isola 2000	Valberg	Estenc	Auron	Mean
SP1 (129 days)	2.2	3.2	1.6	2.7	2.1	2.3
SP2 (79 days)	14.6	16.6	15.6	19.8	17.8	16.9
SP3 (35 days)	38.1	37	34.4	39.3	41.3	38

3.2 Selection of the atmospherics variables fields

3.2.1 Tests of correlation

The Table 3 represents the r between the AV and the 24-hour SP of the WSN from 1997 to 2019. The 18:00 UTC time step always obtained the highest r score, as the pixel closest to the station, and it has been found that all the statistical relations follow a Gaussian function. Thus, according to the work of Baur, a r value of 0.122 shows significant relations. The majority of the covariations (except: specific humidity 850 hPa; Air temperature 500 hPa; U-Component-Wind 500 and 700 hPa; V-Component-Wind 850 hPa; Vertical velocity 500 and 850 hPa; 2 m air temperature; Sea surface temperature) are above this threshold. However, the intensity of the relations is not homogeneous. A choice in the AV was made, according to the value of the coefficient r (the most intense relations are preserved) and the absence of redundancy between V (for instance, the surface pressure is discarded since it is strongly correlated with the 700 hPa geopotential, whose relationship with the 24-hour SP is superior). In accordance with these criteria, five AV have been chosen:

Table 3

Results of the correlation (r) tests.
AVs	Levels	Time steps	ISOLA	AURON	VALBERG	ESTENC	VAL-C.	WSN
Geopotential	500 hPa	18h UTC	-0,27	-0,24	-0,19	-0,28	-0,24	-0,24
Geopotential	700 hPa	18h UTC	-0,31	-0,27	-0,29	-0,33	-0,27	-0,29
Geopotential	850 hPa	18h UTC	-0,28	-0,26	-0,29	-0,34	-0,25	-0,28
Specific humidity	500 hPa	18h UTC	0,22	0,27	0,23	0,25	0,18	0,23
Specific humidity	700 hPa	18h UTC	0,30	0,31	0,35	0,32	0,26	0,31
Specific humidity	850 hPa	18h UTC	0,08	0,09	0,09	0,13	0,05	0,09
Air temperature	500 hPa	18h UTC	-0,11	-0,05	-0,08	-0,07	-0,10	-0,08
Air temperature	700 hPa	18h UTC	-0,17	-0,14	-0,15	-0,14	-0,15	-0,15
Air temperature	850 hPa	18h UTC	-0,25	-0,22	-0,22	-0,22	-0,23	-0,23
U-Component-Wind	500 hPa	18h UTC	0,02	0,12	0,09	0,15	0,02	0,08
U-Component-Wind	700 hPa	18h UTC	-0,08	-0,03	-0,02	0,03	-0,07	-0,03
U-Component-Wind	850 hPa	18h UTC	-0,26	-0,27	-0,25	-0,16	-0,19	-0,23
V-Component-Wind	500 hPa	18h UTC	0,37	0,37	0,36	0,36	0,35	0,36
V-Component-Wind	700 hPa	18h UTC	0,33	0,43	0,41	0,42	0,37	0,39
V-Component-Wind	850 hPa	18h UTC	-0,08	0,25	0,14	0,10	-0,08	0,07
Vertical velocity	500 hPa	18h UTC	-0,08	-0,11	-0,12	-0,13	-0,09	-0,11
Vertical velocity	700 hPa	18h UTC	-0,23	-0,28	-0,29	-0,32	-0,35	-0,29
Vertical velocity	850 hPa	18h UTC	-0,15	-0,02	-0,04	0,04	0,20	0,01
Surface pressure	Surface	18h UTC	-0,29	-0,27	-0,29	-0,34	-0,24	-0,29
M.S level pressure	Surface	18h UTC	-0,23	-0,20	-0,25	-0,32	-0,20	-0,24
2 m air temperature	Surface	18h UTC	-0,09	-0,07	-0,11	-0,01	-0,16	-0,09
Sea temperature	Surface	18h UTC	-0,08	-0,02	-0,01	0,04	-0,07	-0,03

Vertical velocity 700 hPa (WSN r = -0.29). It is the engine that produces the precipitation mechanism. It can be associated with several phenomena: thermal instability, relief, low pressure system.
Geopotential height 700 hPa (WSN r = -0.29). The geopotential makes it possible to guess the center of actions (low pressure or high pressure) and implicitly the lift or the subsidence of the air mass.
Specific humidity 700 hPa (WSN r = 0.31). It is the water content of the air mass, therefore the raw material of precipitation.
Air temperature 850 hPa (WSN r = -0.23). This physical quantity is directly linked to the production of precipitation in solid form. Indeed, sufficiently low temperatures ( = < 0°C) are necessary for flakes to form.
V component of the wind 700 hPa (WSN r = 0.39). This is the Southern direction of the wind. This variable is highly correlated with the 24-hour SP and shows the importance of the origin of the air mass in the mechanism of production.

3.2.2 Mean configurations of the atmospherics variables per classes of solid precipitations

The role of the AV in the SP production mechanism is different. The geopotential height 700 hPa, the V component of the wind and the vertical velocity, act on the movement of the air mass. The specific humidity partly controls the intensity of the SP, since it is characterized as the raw material. Finally, the air temperature affects the state of the precipitation, which becomes solid if the atmosphere is equal or below 0°C. The mean configurations per classes of SP of the geopotential height 700 hPa (cf. Figure 6), specific humidity 700 hPa (cf. Figure 7) and air temperature 850 hPa (cf. Figure 8) are presented, while those of the vertical velocity 700 hPa and the V component of the wind 700 hPa are respectively available in Fig.S1 and Fig.S2.

The Fig. 6 present a southwest cyclonic circulation, with a trough centered from Great Britain to the Mediterranean. This trough is accentuated when passing from one class to another. The specific humidity 700 hPa (Cf. Figure 7), shows a North of Europe relatively dry, with around 1.2 to 1.5 g/kg of moist air. On the other hand, in the South of France, the perimeter of the Mediterranean basin is humid, with values oscillating between 1.8 and 2.4 g/kg of moist air, in the case of the most intense events (SP3). Finally, the air temperature 850 hPa, indicates 0°C close to the WSN (Cf. Figure 8). In the case of SP2 and SP3, this cold air is surrounded by warmer air, which could also strengthen the convection mechanism of the air mass from the Mediterranean Sea, by a confrontation between cold air and warm air, synonymous with thermal instability.

The mean configurations of the predictors per classes of SP were presented. The next part is interested in the results of the classification of the AV fields. How well do these mean configurations per classes of SP, match with the mean configurations of the ACT?

3.3 Classification of the atmospherics variables fields

This part concerns the result of the classification computed by PCA and HCA on the five AV fields provided by the ERA5 reanalysis and chosen according to the correlation tests (cf. Part 3.2.1). The main objective is to obtain ACT specific to the SP. As a reminder, two classifications were carried out, a first from 1997 to 2019 (DJFMA) which serves as a control classification, then a second from 1950 to 2019 (DJFMA), which makes it possible to analyze the evolution of atmospheric conditions specific to SP over a long period of time (70 years).

3.3.1 Principal Component Analyze on the atmospherics variables fields

The PCA covers 70 years (1950–2019 DJFMA) and has a total of 10 708 days. The five data matrix (1 785 grid points x 10 708 days for each AV) is subjected to a PCA. According to the rule of Kaiser 67 PC are retained (Cf. Table 4) and form a new matrix of 67 columns and 10 708 rows. This matrix will be used in the HCA. It is usually preferable to keep a larger number of axes to cover at least 80 to 90% of the variance (Martin & Saintignon, 1974). Thus, for the five variables this threshold is respected.

Table 4

Number of principal components selected (Kaiser’s rule) in the PCA.
AV	Number of PC	% of the variance
Geopotential height 700 hPa	9	83
S. Humidity 700 hPa	16	91
Air temperature 850 hPa	12	84
V component of W. 700 hPa	11	89
Vertical velocity 700 hPa	19	92
Total	67

3.3.2 Hierarchical Clustering Analyze on the atmospherics variables fields

Following the PCA a new table of data is obtained, and it contains the 67 selected components in columns and the 10 708 days (1950–2019 DJFMA) in rows. It is therefore a question of discretizing the 10 708 days into several classes of synoptic conditions, or in other words, it is a question of producing the ACT. Indeed, an HCA is carried out according to the method of Ward. The choice of the number of classes is established following the dendrogram (cf. Figure 9) and the jumps of inertia (cf. Figure 10), in total, six classes are retained. There is no single ideal level of cut of the hierarchical tree, because the different levels correspond to an increasingly detailed point of view of the grouping. Nevertheless, when the inertia jump on the axis from one grouping to another is strong, it is considered that the separation of the groups is all the better (Chesnezu, 2021). There is a final jump in inertia between class 6 and class 7, which gives credibility to the chosen grouping.

3.3.3 Mean configurations of the atmospheric circulation types

The combination of the PCA and the HCA gave six classes (cf. Table 5) of days with similar atmospheric circulation patterns (ACT). The number of days per ACT class is quite similar from one group to another [1522; 2158]. ACT5 is the one with the most days (2158), while ACT 6 is the one with the fewest (1522).

Table 5

Synthesis of the classification of reanalysis fields.
ID	Type of circulation	Number of days (1950–2019 DJFMA)	Percent of days
ACT1	Depression over Central Europe (DCE)	1582	15
ACT2	Northwest anticyclonic (NWA)	1818	17
ACT3	South, trough on Western Europe (STW)	1815	17
ACT4	West cyclonic (WC)	1813	17
ACT5	West anticyclonic circulation (WA)	2158	20
ACT6	Southwest anticyclonic (SWA)	1522	14

The mean configurations (average of all days belonging to the class) are presented for the six ACT. These maps can be compared to the mean configuration of the predictors per classes of SP (Cf. Figures 6, 7 and 8), to see if the classification has succeeded in the reproduction of an ACT specific to SP and will make it possible to qualify the reversibility of the approach. The mean configurations per classes of ACT of the geopotential height 700 hPa (Cf. Figure 11), specific humidity 700 hPa (Cf. Figure 12) and air temperature 850 hPa (Cf. Figure 13) are very demonstrative (those of the vertical velocity 700 hPa and the V component of the wind 700 hPa are respectively available in Fig.S3 and Fig.S4).

The six ACT cover all the synoptic configurations, from the low-pressure situation to the high blocking over the whole of Europe. The ACT1 appears as a depression over central Europe (DCE), characterized by low altitudes of the geopotential height 700 hPa (values are close to 2800 meters over the WSN) and a low-pressure trough, which extends from the Baltic Sea to Corsica. The air is dry on the North of Europe and the Mediterranean basin, with values close to 1.2 g/kg of moist air on the WSN. The air temperature is very cold, with − 2°C to -6°C on France. The ACT2 corresponds to a northwest anticyclonic (NWA) circulation, with a ridge to the west of France, which extends from the Canary Islands to Brittany and a low-pressure trough to the East. The specific humidity is low with 0.9 to 1.2 g/kg of moist air on the WSN. The air is cold in France, with temperature between 0°C to -4°C. The ACT3 is a southern cyclonic circulation with a trough on western Europe (STW). It is marked by a low-pressure area over the British islands and a low-pressure trough centered northwest-southeast, going from the British islands to the Balearic Islands. The direction of the flow is SW, with a high level of specific humidity (2.1 to 2.4 g/kg of moist air throughout the Mediterranean basin). The air temperature is close to 0°C on the WSN, surrounded by slightly higher values of 2°C. The ACT4 has the characteristics of a West cyclonic (WC) circulation, with a rapid flow in altitude, as shown by the tightening of the isohypses. The specific humidity and the air temperature are medium over France. The ACT5 is a West anticyclonic (WA) circulation, with an anticyclone centered on the Balearic Islands and a powerful ridge centered southwest-northeast, which extends from the Maghreb to Switzerland. The anticyclonic blocking situation leads to an almost total cut off, of the West circulation over Europe, the air here is therefore stable, warm (6°C above the WSN) and dry (0.9 g/kg of moist air above the WSN). Finally, the ACT6 is understood as a southwest anticyclonic (SWA) circulation, whose anticyclonic ridge extends from Libya to Hungary. The humidity is quite high, with values around 2.4 to 2.7 g/kg of moist air on the South of France. The air temperature is warm with 4°C to 6°C above the WSM.

3.4 Association between the classes of solid precipitations and the ACT

3.4.1 Similarity in circulation patterns

The comparison between the mean configurations per class of SP (cf. Part 3.2.2) and the mean configuration per ACT (cf. Part 3.2.3), shows that the STW correctly reproduces the circulation patterns in case of snowy days. Indeed, in both cases, a South cyclonic circulation with a trough over western Europe is observed. Table 6 presents the mean deviation values for the geopotential height 700 hPa and the specific humidity 700 hPa, between the mean configurations for the three classes of SP and the mean configuration for the STW. The comparison is made on these two variables since they represent the transport of the air mass and the raw material of the SP. The evaluation is done on a widened zone around the WSN (cf. Fig.S5) and show small deviations between the datasets (6 meters for the geopotential and 0.12 g/kg for the specific humidity). Also, it is found that the 1950–2019 classification is more efficient (lower difference) than the control classification (1997–2019). It seems like the longer temporal extension of the 1950–2019 classification brings more stability to the ACT produced.

Table 6

Evaluation of the deviations for the geopotential and the specific humidity between the mean configurations by SP classes and the mean configuration of STW.
Type of configurations	Period	Geopotential 700 hPa	Specific humidity 700 hPa
SP (mean of SP1, SP2 and SP3)	1997–2019	2937 hPa	2.18 g/kg
STW (classification control)	1997–2019	2947 hPa	2.04 g/kg
STW	1950–2019	2943 hPa	2.3 g/kg

3.4.2 Association between the classes of solid precipitations and the ACT

The Fig. 14 presents the 243 SP, associated with each ACT, between 1997 and 2019, and the Fig. 15 corresponds to the total amount of SP (fell in 24 hours) associated with each ACT. These two figures are the second criteria to validate the efficiency of the classification specific to SP. It is clearly perceived the role of the STW, which leads 150 SP (62%) and an amount of 2 562 cm (out of 3 285 cm in total). The STW is also associated with 72% of the SP2 and 100% of the SP3. DCE is the second most productive class, with 59 SP (24%), and an amount of 459 cm. NWA and WC, are not very productive with respectively 13 and 21 SP (amount of less than 190 cm). Finally, WA and SWA produce no SP. These two classes are the opposite extreme of STW, since they offer some synoptic conditions very unfavorable to the SP.

There is therefore a statistical link between STW and the 24-hour SP, as shown in Fig. 16, which represents the seasonal number of days of occurrence (orange curve) of STW and the seasonal amount of snow caused by the 24-hour SP (blue curve) at the Isola 2000 weather station (Isola 2000 is the only station of the WSN to have some SP measured since 1985). The two curves follow the same dynamic, for example, the decrease in the seasonal amount of SP observed at Isola 2000 between the 1990s, corresponds to the decrease in the occurrences of STW. The r has a score of 0.75 (the two series of data have a Gaussian function), which shows a positive linear relationship between the two variables.

3.5 Evolution of the occurrence of the atmospheric circulation types

3.5.1 Frequency of the atmospheric circulation types

The DCE, but mainly the STW, are the synoptic circulations that produced a large majority of the 24-hour SP. Conversely, the WA and SWA did not produce any SP events. The multi-decadal evolution of the total number of days per season of these four classes was calculated (cf. Figure 17). The Mann-Kendall (MK) test is applied to each series to ensure the significance of the results (Mann 1945, Kendall 1975, Gilbert 1987) and all the linear trends presented below validated this test.

The multi-decadal evolution of DCE, shows a decrease during the decades 1950–1959, 1960–1969, 1970–1979, 1980–1989 and 1990–1999, with a minimum of 114 days for 1990–1999. There is a slight increase for the decades 2000–2009 and 2010–2019. However, the 70-year trend remains largely downward as illustrated by the trend line. The multi-decadal evolution of STW, shows an increase during the decades 1950–1959, 1960–1969 and 1970–1979, with a maximum of 346 days for 1970–1979. STW follows a sharp drop for the decades 1980–1989, 1990–1999, 2000–2009 and 2010–2019 with a minimum of 210 days for 1990–1999. The 70-year trend is downward. The WA follows the reverse logic, with a steady decrease for the decades 1950–1959, 1960–1969 and 1970–1979, which drops to a minimum of 218 days. Then, it follows an increase for the decades 1980–1989, 1990–1999, 2000–2009 and 2010–2019, with a maximum of 415 days for 1990–1999. Thus, the 70-year trend shows an increase in its occurrence, as illustrated by the trend line. Finally, the SWA shows a clear increase during the period 1960–2019.

3.5.2 Internal variability of the atmospheric circulation types

The internal variability of the ACT corresponds to the persistence in number of consecutive days of the class when it is established. This is a second indicator which makes it possible to observe the weakening or not of a class over time. This variability is calculated for the four ACT, then an MK test is applied. Only WA showed a significant positive trend (cf. Figure 18). This class has strengthened over the past 70 years, going from an average of 3 consecutive days to around 4 consecutive days.

3.5.3 Evolution of the air temperature 850 hPa in case of STW

The seasonal evolution of the air temperature 850 hPa in case of STW (mean values taken above all the pixels ERA5 of the WSN), allows to observe inside the main SP circulation class, the importance of the warming of the air. Is the temperature during synoptic conditions conducive to SP increasing? This question is essential, since if the rise in temperature comes to impact STW, then the formation process of snow is threatened, since the efficiency of the synoptic class is called into question. To answer this question, the mean temperature per season at 850 hPa (cf. Figure 19) and the mean temperature during STW (cf. Figure 20) are compared.

The mean temperature per season (1950–2019 DJFMA) at 850 hPa shows a strong upward trend (with an accentuation since the 1985s), going from 2°C per season at the beginning of the 1950s to more than 3°C for the years 2010. This trend is validated by the MK test. The mean temperature during STW also shows an upward trend (validated by the MK test). The mean seasonal temperature is around 0.8°C in 1950, then reaches 1°C in 1980 and finishes to 1.6°C in 2019. It is noted an increase of + 0.8°C in the space of 70 years, and two peak years (2012 with 3.2°C and 2019 with 2.8°C) occurred in the last decade (2010–2019). In the case of STW, the positive trend remains less intense, but suggests that global warming also affects the atmospheric conditions conducive to SP.

4.1 Robustness of the classification

The classification (DJFMA, 1950–2019) of the atmospheric circulation specific to the SP computed six different ACT, including two synoptic classes of SP (STW and DCE), and two dry classes (WA and SWA). It is clearly perceived the role of the STW in the occurrence of the 24-hour SP, because it leads 150 SP (62%) and an amount of 2 562 cm (out of 3 285 cm in total). The STW is also associated with 72% of the SP2 and 100% of the SP3. In addition, the comparison between the patterns of circulation per class of SP, with them for STW, shows small deviations between the two datasets around the WSN, with only 6 meters for the geopotential height 700 hPa and 0.12 g/kg for the specific humidity 700 hPa. It is not surprising to observe a statistical link between the occurrences of STW and the 24-hour SP (r = 0.75). Thus, the classification reproduces well the atmospheric conditions that lead to the SP, especially for the SP2 and SP3, which are the most important events in the recharge of the natural snow cover for human economic activities (damage for ski resorts). The method used is reliable and reproducible for other geographical areas since it is based on local statistical relationships. Finally, the atmospheric data used is free and can easily be downloaded from the ECMWF's climate data store portal.

4.2 Statistical link with the atmospheric variables

The classification is built on the statistical links between five AV and the 24-hour SP over the DJFMA 1997–2019 period, by using some correlation tests, then it is computed and studied over the period 1950–2019. So, it is assumed that the statistical links for the period 1997–2019 are unchanged and remains the same for the previous period 1950–1996. To ensure the significance of the 1950–2019 classification, a control classification was first carried out over the 1997–2019 period. The latter served as a model and made it possible to observe that the results remained robust in both cases. It was even found that the 1950–2019 classification had lower difference in the deviation in the patterns of circulation than the control classification. The temporal extension is longer and brings more stability to the ACT produced. Ideally, it would have been wiser to carry out correlation tests and classification over the entire period 1950–2019, unfortunately the WSN did not allow for such a long common study period.

4.3 Evolution of the patterns of circulation and climate in the South of France

The evolution of the seasonal occurrence of STW and DCE showed a negative trend, while WA and SWA obtained a positive trend. These developments are worrying for the future of the snow and for the climate in the South of France, which seems to be heading towards an increase in anticyclonic days.

A certain natural variability seems to be behind this modification, as already evoked in the scientific literature. For example, Capozzi et al. (2022), explained that the reverse trend of the Arctic Oscillation (AO) index and the prevalence of a neutral or negative Eastern Atlantic Western Russia (EAWR) pattern follow natural cycle and played an important role in the frequency of the SP in the Italian Apennine mountains.

However, in the case of the South of France, the seasonal variability of the ACT is weakly explained by the annular mode of variability. Indeed, correlation tests conducted between ACT and North Atlantic Oscillation (NAO) or AO, gave low r coefficients (NAO r = 0.15, AO r = 0.26). Another explanation for this variability could be related to the Hadley circulation. In fact, the latest IPCC report (Chap. 8 of the 6th IPCC report, 2021) discusses aspects of large-scale atmospheric circulation that have likely changed since the mid-20th century. A part of this chapter concerns the Hadley circulation which has very probably broadened since at least the 1980s (Fig. 2.17 of the report), and the tracks of extratropical storms have probably shifted towards the poles in both hemispheres. This phenomenon could lead to a move further north of the Azores anticyclone, which would therefore be more present at the latitude of the WSN and would have the impact of increasing the occurrence of WA and SWA and decreasing STW and DCE.

Conclusion

This study focused on the relationship between the 24-hour Solid Precipitation (SP) measured in five Alpines stations of the Mediterranean region (French department of the Alpes-Maritimes), for the months of December, January, February, March, and (DJFMA), and the hourly Atmospheric Variables (AV), coming from the ERA5 reanalyses of the European Centre for Medium-Range Weather Forecasts (ECMWF). The 24-hour SP were classified into 3 classes of intensity: SP1 (2.6 cm), SP2 (16.9 cm), SP3 (38 cm).

The main objective was to obtain some Atmospheric Circulation Types (ACT - group of days with a similar atmospheric circulation pattern) reproducing the synoptic conditions causing the SP. The secondary objective was to study statistically the occurrence of the ACT, to define whether the synoptic conditions favorable to SP have evolved between 1950 and 2019. The classification computed (by a combination of a PCA and HCA) six ACT, including two SP classes (STW and DCE) and two dry classes (WA and SWA). Over the period 1997–2019 DJFMA, the South circulation with a trough on Western Europe (STW) has resulted 150 SP (62% in total) and worked especially well on the SP2 (72%) and the SP3 (100%). The comparison between the patterns of circulation per classes of SP, with them for STW, shows strong similarities and demonstrate the success of the approach. The depression over central Europe (DCE) produced 59 SP (24% in total). Conversely, the West anticyclonic (WA) and a southwest anticyclonic (SWA) circulations did not produce any SP events.

The total seasonal number of days per decade of STW and DCE, shows a decrease for the period 1950 to 2019, while WA and SWA, indicates the opposite trends, with an increase for the same period. In addition, WA, have strengthened over the past 70 years, going from an average of 3 consecutive days, for around 4 consecutive days. Thus, synoptic configurations conducive to SP seem to be less frequent in favor of anticyclonic atmospheric configurations producing dry weather in the region. Part of the explanation for this phenomenon could be found in the spreading of the Hadley cell (which has very probably broadened since at least the 1980s) which pushes the Azores anticyclone further North. The Azores anticyclone rises more often in the latitude of southern France and favors the appearance of WA and SWA. Moreover, the increase in temperatures also threatens the SP in the Southern Alps. The mean temperature during STW shows an upward trend of + 0.8°C in the space of 70 years. This result suggests that global warming also affects the effectiveness of the STW in the SP production cycle, which could further lead to precipitation in the liquid form.

Finally, they are serious concerns for the region with a real economic risk for the inhabitants of mountain territories, but also impacts on the hydrological cycle and even more a potential worsening of periods of drought. This classification universal and applicable to other area, will be completed for the future period. The same method will be produced with the AV coming from the data of the six^th Coupled Model Intercomparison Project (CMIP6), to obtain the probable evolutions of the DCE, STW, WA and SWA classes until 2100.

AV: Atmospheric Variables

ACT: Atmospheric Circulation Types

CMIP: Coupled Model Intercomparison Project

GHG: Greenhouse Gas

HCA: Hierarchical Clustering Analyze

PCA: Principal Component Analyze

SP: Solid Precipitation

SSPs: Shared Socioeconomic Pathways

WSN: Weather Station Network

Units

°C: degrees Celsius

g/kg: gram per kilogram

hPa: hectoPascal

m: meters

%: percent

Acknowledgements

I thank the French national meteorological service “Météo-France” to provide me observations datas.

I thank my research laboratory “ESPACE” to let me publish this article.

Formatting of funding sources

This work was not supported by any funding sources.

Author Declarations

Competing Interests

The authors have no relevant financial or non-financial interests to disclose.

Ethics approval

The authors have no ethics approval to declare.

Consent to participate

Not applicable.

Consent for publication

Not applicable.

Author Contributions

All authors contributed to the study conception and design. Treatment was completed by Guillaume Guerin. The first draft of the manuscript was written by Guillaume Guerin and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.

Data Availability

The datasets generated during and/or analysed during the current study are available here:

ERA5 reanalysis: https://cds.climate.copernicus.eu/cdsapp#!/search?type=dataset
24-hour solid precipitation: https://publitheque.meteo.fr

Code availability

Not applicable.

Ashcroft L., Coll J-R., Gilabert A., Domonkos P., Brunet M., Aguilar E., Castella M., Sigro J., Harris I., Unden P., Jones P., 2018. A rescued dataset of sub-daily meteorological observations for Europe and the southern Mediterranean region, 1877–2012, Earth Syst. Sci. Data, 10, 1613–1635, https://doi.org/10.5194/essd-10-1613-2018.
Barbier B., 1982. L'impact des activités humaines (sports d'hiver) sur le milieu montagnard (Control of the mountain ecosystem by winter-sports). Bulletin de l'Association de Géographes Français 485–486 pp. 121–124.
Barrier N., Tréguier A-N., Cassou C., Deshayes J., 2014. Influence des régimes de temps atmosphériques sur la circulation océanique de l'Atlantique Nord. La Météorologie, Météo et Climat, 2014, 8 (87), pp.38–44.
Barry R., & Gan T., 2022. Snowfall and snow cover. In The Global Cryosphere: Past, Present, and Future (pp. 15–86). Cambridge: Cambridge University Press.
Beniston M., 2012. Climate Change and Impacts: From Global to Local. Advances in Global Change Research (AGLO, volume 19). ISBN: 978-1-4020-2346-0.
Berard-Chenu L., François H., Morin S., George E., 2022. The deployment of snowmaking in the French ski tourism industry: a path development approach, Current Issues in Tourism, DOI: 10.1080/13683500.2022.2151876.
Boé J., Terray L., 2008. Régimes de temps et désagrégation d’échelle. La Houille Blanche, N°2 (Avril 2008), pp. 45–51.
Bokhorst S., Pedersen S.H., Brucker L., 2016. Changing Arctic snow cover: A review of recent developments and assessment of future needs for observations, modelling, and impacts. Ambio 45, 516–537 (2016).
Capozzi V., De Vivo C., Budillon G., 2022. Synoptic control over winter snowfall variability observed in a remote site of Apennine Mountains (Italy), 1884–2015. The Cryosphere, 16, 1741–1763, 2022, https://doi.org/10.5194/tc-16-1741-2022.
Carrega P., 2004. « Avant-propos sur les « types de temps » », Norois, 191 | 2004, 7–9.
Carrega, P., et al., 2021. Une catastrophe hors norme d’origine m ́et ́eorologique le 2 octobre 2020 dans les montagnes des Alpes-Maritimes. Physio-G ́eo 16, 70. https://doi.org/10.4000/physio-geo.12370.
Chiadjeu O., Gratton Y., St-Hilaire A., 2016. Analyse multivariée des paramètres morphologiques des lacs de la province du Québec. Rapport No R1684, INRS-ETE, Québec (QC): vi + 23 p.
Chochon, R., Martin, N., Lebourg, T., 2021. Analysis of extreme precipitation during the Mediterranean event associated with the Alex storm in the Alpes-Maritimes: atmospheric mechanisms and resulting rainfall. In: SimHydro 2021 – Models for Complex and Global Water Issues-Practices and Expectations. Sophia-Antipolis France, pp. 16–18. June 2021.
Creusier J., Biétry F., 2014. Analyse comparative des méthodes de classifications. L'exemple du bien-être au travail. RIMHE: Revue Interdisciplinaire Management, Homme & Entreprise, 2014/1 (n° 10, vol. 3), p. 105–123. DOI : 10.3917/rimhe.010.0105. URL : https://www.cairn.info/revue-rimhe-2014-1-page-105.htm.
Croce P., Formichi P., Landi P., Mercogliano P., Bucchignani E., Dosio A., Dimova S., 2018. The snow load in Europe and the climate change, Climate Risk Management, Volume 20, 2018, Pages 138–154, ISSN 2212 – 0963.
Esteban P., Jones P.D., Martín-Vide J. and Mases M., 2005. Atmospheric circulation patterns related to heavy snowfall days in Andorra, Pyrenees. Int. J. Climatol., 25: 319–329.
Franco C., 2021. Environmental History, The History of Tourism in the Mountains and the Construction of New Knowledge: A Study of the Architecture of Winter Sports Resorts in the French-Italian Alps », Journal of Alpine Research | Revue de géographie alpine.
François H., Morin S., Spandre P., Lafaysse M., Lejeune Y., George-Marcelpoil E., 2016. Croisement de simulations numériques des conditions d’enneigement avec une base de données socio–économiques spatialisée des stations de sports d’hiver: description de l’approche, application aux Alpes françaises et introduction de la prise en compte des pratiques de gestion (damage et neige de culture). La Houille Blanche, n° 4, 2016, p. 66–84.
Guerin G., Martin N., 2022. Atmospheric circulation conditions and snowy weather in the Southern Alps: Case of the Isola 2000 weather station. Polar Science. 10.1016/j.polar.2022.100922.
Gilbert R.O., 1987. Statistical Methods for Environmental Pollution Monitoring, Wiley, NY.
Hanzer F., Carmagnola C-M., Ebner P-P., Koch F., Monti F., Bavay M., Bernhardt M., Lafaysse M., Lehning M., Strasser U., François H., Morin S., 2020. Simulation of snow management in Alpine ski resorts using three different snow models. Cold Regions Science and Technology, Volume 172, 102995, ISSN 0165–232X, https://doi.org/10.1016/j.coldregions.2020.102995.
IPCC, 2021: Climate Change 2021. The Physical Science Basis. Contribution of Working Group I to the Sixth Assessment Report of the Intergovernmental Panel on Climate Change [Masson-Delmotte V., P. Zhai A., Pirani S.L., Connors C., Péan S., Berger N., Caud Y., Chen L., Goldfarb M.I., Gomis M., Huang K., Leitzell E., Lonnoy J.B.R., Matthews T.K., Maycock T., Waterfield O., YelekçI R., Zhou B. Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA.
Irannezhad M., Ronkanen A-K., Kiani S., Chen D., and Kløve B., 2017. Long-term variability and trends in annual snowfall/total precipitation ratio in Finland and the role of atmospheric circulation patterns. Cold Reg. Sci. Technol., 143, 23–31, https://doi.org/10.1016/j.coldregions.2017.08.008, 2017.
Joye J-F., 2019. Enneigement artificiel, évaluation environnementale des projets d’aménagements touristiques en montagne. Le point après l’annulation partielle du décret UTN. Journal Spécial des Sociétés.
Kendall M.G., 1975. Rank Correlation Methods, 4th edition, Charles Griffin, London.
Klein G., 2018. Variabilité du manteau neigeux des Alpes Européennes entre 1950 et 2016 dans un contexte de changement climatique: revue bibliographique. Climatologie 15 22–45 (2018).
Kyselý, J., Beguería, S., Beranov ́a, R., Ga ́al, L., Lo ́pez-Moreno, J.-I., 2012. Different patterns of climate change scenarios for short-term and multi-day precipitation extremes in the Mediterranean. Global Planet. Change Volumes 98–99, 63–72.
Lastrada E., Garzón-Roca J., Cobos G., Torrijo FJ., 2021. A Decrease in the Regulatory Effect of Snow-Related Phenomena in Spanish Mountain Areas Due to Climate Change. Water. <background-color:#CFBFB1;uvertical-align:baseline;>https://doi.org/10.3390/w13111550</background-color:#CFBFB1;uvertical-align:baseline;>
Lavers D. A., Harrigan S., & Prudhomme C., 2021. Precipitation Biases in the ECMWF Integrated Forecasting System, Journal of Hydrometeorology, 22(5), 1187–1198.
Mahto S-S., Mishra V., 2019. Does ERA-5 outperform other reanalysis products for hydrologic applications in India? Journal of Geophysical Research: Atmospheres, 124. https://doi.org/10.1029/2019JD031155
Mann H.B., 1945. Non-parametric tests against trend, Econometrica 13:163–171.
Marnezy A., 2008. Alpine dams. Journal of Alpine Research | Revue de géographie alpine [Online], 96 – 1 | 2008. URL: http://journals.openedition.org/rga/430; DOI: https://doi.org/10.4000/rga.430.
Martin N., 2008. La pollution par l'ozone et la climatologie dans un espace méditerranéen: les Alpes-Maritimes. Géographie. Université Nice Sophia Antipolis, 2008. Français.
Martin N., Carrega P., 2011. L’enneigement dans les Alpes du sud en France: évolution temporelle et relation avec la circulation atmosphérique, Actes du XXIVème colloque AIC, Rovereto septembre 2011, pp 387–392.
Martin S., Saintignon M-F., 1974. Application de l'analyse factorielle en composantes principales à l'étude de la variabilité régionale et interannuelle des précipitations ardéchoises. In: Revue de géographie de Lyon, vol. 49, n°1, 1974. pp. 77–92. DOI: https://doi.org/10.3406/geoca.1974.1638.
Merino A., Fernández S., Hermida L., López L., Sánchez JL., García-Ortega E., Gascón E., 2014. Snowfall in the northwest Iberian Peninsula: synoptic circulation patterns and their influence on snow day trends. ScientificWorldJournal. doi: 10.1155/2014/480275. Epub 2014 Jul 24. PMID: 25152912; PMCID: PMC4134785.
Morin S., François H., Réveillet M., Sauquet E., Crochemore L., Branger F., Leblois E., Dumont M., 2022. Quantifying the hydrological disturbances induced by snow grooming and snowmaking in ski resorts: a case study in the French Alps, EGUsphere [preprint], https://doi.org/10.5194/egusphere-2022-1186, 2022.
Nogueira M., 2020. Inter-comparison of ERA-5, ERA-interim and GPCP rainfall over the last 40 years: Process-based analysis of systematic and random differences. Journal of Hydrology, Volume 583.
Nunnally J. C., Bernstein I-H., 1994. Psychometric Theory. New York, McGraw-Hill.
Parker W., 2016. Reanalyses and Observations: What’s the Difference?. Bulletin of the American Meteorological Society. 97. 160128144638003. 10.1175/BAMS-D-14-00226.1.
Pathak R., Sahany S., Mishra S.K. et al., 2019. Precipitation Biases in CMIP5 Models over the South Asian Region. Sci Rep 9, 9589 (2019).
Pédelaborde P., 1956. La circulation générale de l'atmosphère. L'information géographique, volume 20, n°3.
Peña J-C., Schulte L., 2019. Simulated and reconstructed atmospheric variability and their relation with large Pre-industrial summer floods in the Hasli-Aare catchment (Swiss Alps) since 1300 CE. Global and Planetary Change, Volume 190.
Pettitt A-N., 1979. A Non-Parametric Approach to the Change-Point Problem. Journal of the Royal Society: Series C (Applied Statistics), 28, 126–135.
https://doi.org/10.2307/2346729.
Piras, M., Mascaro, G., Deidda, R., Vivoni, E., 2016. Impacts of climate change on precipitation and discharge extremes through the use of statistical downscaling approaches in a Mediterranean basin. Sci. Total Environ. 543.
Planchon O., Cantat O., Quénol H., 2015. Types de temps et types de circulation atmopshérique: Essai de mise en relation à Angers et Bordeaux (France). XXVIIIe Colloque de l’Association Internationale de Climatologie, Liège 2015.
Poccard I., 2000. Étude diagnostique de nouvelles données climatiques: les réanalyses.
Exemples d'application aux précipitations en Afrique tropicale. Géographie. Université de Bourgogne. Français.
Silber I., Verlinde J., Wang S-H., Bromwich D., Fridlind A., Cadeddu M-P., Eloranta E., Flynn C., 2019. Cloud Influence on ERA5 and AMPS Surface Downwelling Longwave Radiation Biases in West Antarctica. Journal of Climate. 32. 10.1175/JCLI-D-19-0149.1.
Smith C. D., Ross A., Kochendorfer J., Earle M. E., Wolff M., Buisán S., Roulet Y.-A., and Laine T., 2020. Evaluation of the WMO Solid Precipitation Intercomparison Experiment (SPICE) transfer functions for adjusting the wind bias in solid precipitation measurements, Hydrol. Earth Syst. Sci., 24, 4025–4043.
Stoffel M., Corona C., 2018. Future winters glimpsed in the Alps. Nature Geosci 11, 458–460. https://doi.org/10.1038/s41561-018-0177-6
Viceto C., Gorodetskaya I., Rinke A., Maturilli M., Crewell S., 2021. Atmospheric rivers and associated precipitation patterns during the ACLOUD/PASCAL campaigns near Svalbard (May–June 2017): case studies using observations, reanalyses, and a regional climate model. 10.5194/acp-2021-609.
Ye H., Cohen J., Rawlins M., 2013. Discrimination of Solid from Liquid Precipitation over Northern Eurasia Using Surface Atmospheric Conditions, Journal of Hydrometeorology, 14(4), 1345–1355.
Zhang R., Wang H., Fu Q., Rasch PJ., Wang X., 2019. Unraveling driving forces explaining significant reduction in satellite-inferred Arctic surface albedo since the 1980s. Proc Natl Acad Sci U S A. 2019 Nov 26;116(48):23947–23953. doi: 10.1073/pnas.1915258116. Epub 2019 Nov 11. PMID: 31712425; PMCID: PMC6883849.
Zuo H., Alonso-Balmaseda M., Tietsche S., Mogensen K., and Mayer M., 2019. The ECMWF operational ensemble reanalysis–analysis system for ocean and sea ice: a description of the system and assessment The European Centre for Medium-Range Weather Forecasts, Shinfield Rd, Reading RG2 9AX, UK, Ocean science 2019.

No competing interests reported.

Download PDF

Journal Publication

published 24 Nov, 2023

Read the published version in Theoretical and Applied Climatology →

Editorial decision: Major revision
11 Aug, 2023
Reviews received at journal
07 Jun, 2023
Reviewers agreed at journal
07 Jun, 2023
Reviewers agreed at journal
05 Jun, 2023
Reviewers invited by journal
03 Jun, 2023
Editor assigned by journal
02 Jun, 2023
Submission checks completed at journal
02 Jun, 2023
First submitted to journal
01 Jun, 2023

You are reading this latest preprint version

How changements in the circulation patterns specific to the solid precipitations can affect these meteorological events in the Alpine stations of the Mediterranean region ? Use of the ERA5 reanalyses

Status:

Journal Publication

Version 1

Abstract

Figures

1. Introduction

2. Data and methods

2.1 Study area and solid precipitation observations

2.1.1 Data quality control of the 24-hour solid precipitations

2.1.2 Intensity classes of the 24-hour solid precipitations

2.2 Atmospheric variables

2.2.1 Atmospheric circulation types

2.2.2 The ERA5 reanalysis

2.2.3 Selection of the atmospherics variables

2.2.4 Window extraction of the atmospherics variables fields

2.2.5 Method of classification of atmospherics variables fields

3. Results

3.1 Classification of the solid precipitations

3.2 Selection of the atmospherics variables fields

3.2.1 Tests of correlation

3.2.2 Mean configurations of the atmospherics variables per classes of solid precipitations

3.3 Classification of the atmospherics variables fields

3.3.1 Principal Component Analyze on the atmospherics variables fields

3.3.2 Hierarchical Clustering Analyze on the atmospherics variables fields

3.3.3 Mean configurations of the atmospheric circulation types

3.4 Association between the classes of solid precipitations and the ACT

3.4.1 Similarity in circulation patterns

3.4.2 Association between the classes of solid precipitations and the ACT

3.5 Evolution of the occurrence of the atmospheric circulation types

3.5.1 Frequency of the atmospheric circulation types

3.5.2 Internal variability of the atmospheric circulation types

3.5.3 Evolution of the air temperature 850 hPa in case of STW

4. Discussion and interpretation of the results

4.1 Robustness of the classification

4.2 Statistical link with the atmospheric variables

4.3 Evolution of the patterns of circulation and climate in the South of France

Abbreviations

Declarations

References

Additional Declarations

Supplementary Files

Status:

Journal Publication

Version 1