This chapter presents the social, economic and environmental impacts that climate change-induced drought episodes produce for the Paris Metropolitan Area. It presents an estimation of the direct and indirect economic costs induced by future drought episodes.
Adapting the Paris Metropolitan Area to a Water‑Scarce Future
Abstract
Droughts have occurred historically the Paris Metropolitan Area (“the region”), with some recorded as early as the 16th century. The most notable recorded drought took place in 1921, following a very dry winter in 1920 (when there was a 58% precipitation deficit compared with the 1981-2010 average). This drought had major effects on agricultural production, river navigation and ecosystems (Van der Schrier et al., 2021[1]). Other notable droughts include the 1976 drought, which was the result of a major heatwave and low annual precipitation and led to some of the lowest water levels ever recorded in the region (DRIEAT, 2023[2]).
Socio-economic development, such as urban expansion and economic activities have contributed to heightening drought risk in the past. Home to nearly 19% of France’s population, the Paris Metropolitan Area is the most densely populated region in the country (INSEE, 2022[3]). Although trends suggest that population growth is stabilising, the region's population grew by 15% between 1968 and 1990, with 60% of that growth occurring outside of the city’s core (Institut Paris Région, 2021[4]). Urban sprawl is partly responsible for the increase in built-up land, which reduces soil permeability and disrupts groundwater recharge and the natural flow of rainwater towards watercourses, increasing the likelihood of drought. The region's development has also been accompanied by the creation of canals that have altered the natural water cycle and when and where water is available.
Nevertheless, the region has been relatively resilient to the effects of droughts over the past 50 years thanks to an efficient infrastructure network and a drought prevention strategy. Most drought-related damage has been recorded in the agricultural sector and the built environment, which points to the environmental impacts of a lack of water in the soil. Although drinking water supply is under pressure at times, to date the region has maintained sufficient levels to cover potable water demand. Thus far, the region's water supply for sanitation, riverine transport and energy production have not been affected by drought episodes. Maintaining water supply levels was enabled through an effective infrastructure network, a resilient drinking water supply strategy, and an approach to risk reduction based on awareness-raising and regulatory measures. While the 1921 drought would have led to strict restrictions on water use for 151 days in today's context, infrastructure development since then has made it possible to largely avoid imposing restrictions on the use of the region's rivers (EPTB Seine Grands Lacs, 2022[5]).
However, with a marked increase in frequency and intensity of droughts since 2003 drought impacts are starting to be increasingly felt across the region. The duration of droughts increased by 4% on average and the area recorded a fall in soil moisture by 5% between the 1961-90 and 1981-2010 periods (Météo France, 2022[6]). These trends are, mainly concentrated in spring and summer. Climate change projections indicate this trend may continue, suggesting a significant drop in future river flow rates, an increase in the length of droughts and, consequently, a fall in groundwater levels towards the driest levels recorded.
This chapter aims to understand the potential impacts of increasing water scarcity in the Paris metropolitan area as a result of severe droughts, made increasingly likely by climate change, and to assess the robustness of the measures currently in place. It focuses on drought and heatwave hazards that contribute to water scarcity and it establishes likely future scenarios, taking into account the existing measures in the Seine-Normandy river basin to prevent drought-related risks. It also analyses trends in water abstraction and consumption in the region, and its vulnerability to a drop in the resources available. Finally, it presents an economic assessment of the costs that the Paris metropolitan area could face in the event of a severe drought and a model of its cascading impacts to better understand the economic effects on a regional and national scale. This impact assessment will make it possible to target specific levers to improve the region's resilience to water scarcity.
The region’s surface water resources rely on the Seine–Normandy river basin and a number of strategic aquifers. Some 55 000 km of rivers flow through the Seine–Normandy basin, including the Seine and its major tributaries (the Yonne, Marne and Oise), as well as many smaller rivers and streams. The basin also has a dozen strategic aquifers with a variety of geological properties, enabling water to be stored and, in some cases, fed into watercourses. The Seine is 776 km long and empties into the English Channel. It supplies water to a number of regions upstream and downstream of the Paris metropolitan area (Figure 2.1).
Source: Authors.
Precipitation patterns influence the flow rates of the Seine and its tributaries, as well as the level of underground reserves. Annual precipitation levels are slightly lower than in other French river basins (Table 2.1) and the low relief of the basin generates relatively low natural flow rates (AESN&DRIEE, 2016[7]). It is estimated that just 30% of precipitation drains into the river basin and recharges the aquifers,1 compared to 50% for the Rhône-Mediterranean basin. The Seine reaches its maximum flow rate in January–February, when evapotranspiration is at its lowest, and its minimum flow rate in August, when evapotranspiration is at its highest. The Seine's average annual flow rate in Paris is 310 m³/s (PIREN-Seine, 2018[8]). Depending on the time of year and the level of precipitation, this flow rate can vary considerably. For example, in Paris, the average interannual flow rate can reach 550 m³/s if there is high precipitation in winter and 150 m³/s when water levels are low in August,2 i.e. almost three times the flow rate recorded during the 1976 drought (EPTB Seine Grands Lacs, 2022[5]).
Source: OECD based on data from the various water agencies, the French Geological Survey and aufrance; French Geological Survey, Climatologie (brgm.fr); Eau Grand Sud-Ouest,3 Agence de l'Eau Seine-Normandie, Le bassin de la Seine [The Seine Basin], eau-seine-normandie.fr.
Droughts can occur due to precipitation deficit, high temperatures or a combination of both. Since droughts can occur when there is a precipitation deficit compared with recorded averages, they can take place in any season, and their severity mainly depends on how long the precipitation deficit lasts. The longer the deficit lasts, the more widespread the consequences of the drought, affecting soil moisture, river flow rates and groundwater levels (Box 2.1). Very high temperatures also make drought more likely because they increase evaporation (i.e. transformation of liquid water to water vapor) and evapotranspiration (i.e. liquid water transformation and soil moisture and plant transpiration) and dry out the soil. The drought that affected the Paris Metropolitan area in 1921, which lasted almost the entire year, is an example of a drought caused by a significant precipitation deficit during winter (58%). During this drought, the Seine River flow was half its summer average in the 19th century, and the region experienced 12 months of dry soil. Were such a drought to occur today, it is estimated that restrictions would have to be imposed on irrigation, navigation, industrial water use, and water use for gardens and by sporting facilities lasting almost 151 days (EPTB Seine Grands Lacs, 2022[9]) It was exacerbated by summer temperatures that increased the evaporation of water from the soil. In contrast, the summer 2019 drought followed a wet winter, but was driven by extremely hot temperatures.
The Paris metropolitan area has been subject to an increase in evapotranspiration in the past decades. Annual average precipitation in the region has risen slightly since 1961, with no significant seasonal differences. However, some years, such as 1921, 1949, 1976 and 2006, saw significant deficits. At the same time, the average temperature in the region has been rising by an average of 0.3°C per decade since 1959, with increasingly frequent and long-lasting heatwaves (Météo France, 2022[10]). These have led to an increase in average annual evapotranspiration of around 6% between 1960-75 and 2003-18 (Figure 2.2).
Note: Annual values for the Paris metropolitan area correspond to the average evapotranspiration in each season. The annual values for France correspond to the average evapotranspiration of all basins in each season.
Source: Authors' work based on World Bank data.
These trends explain the changing pattern of droughts in the region since the beginning of the 20th century. For example, between 1961-90 and 1991-2020, the Paris metropolitan area recorded a fall in soil moisture of around 5% across the year, mainly concentrated in spring and summer. The duration of droughts increased by 4% on average between the 1961-90 and 1981-2010 periods (Météo France, 2022[6]). The drought events of 2006 and 2011 contributed to the driest winter soil recorded since 1959. The 2003 drought – caused by low spring precipitation and a record heatwave – resulted in crop losses of an estimated 60% (Kapsambelis, 2018[11]). More recently, following a particularly dry winter of 2021-22 precipitation deficits reached up to 41% compared with the 1981-2010 average (Météo Paris, n.d.[12]). This drought was remarkable due to both its length (one year) and its intensity. Almost 20% of rivers running through the region were subject to strict restrictions due to low river flow rates (DRIEAT, 2023[3]). Boats were grouped together at locks. Significant damage to buildings was recorded due to clay shrinkage and swelling.
Urban development has also hindered the renewal of underground resources, contributing to a heightened drought risk. Eighty percent of the land in Paris and the surrounding departments is built-up. Twenty-one percent of land in the region is built-up, which is well above the national average of 9% (Agreste, 2021[13]). Building on land reduces the ability of water to infiltrate the soil and recharge aquifers.
Droughts can lead to water scarcity. In the event of a drought characterised by lower flow rates, or below-average groundwater levels, there may not be sufficient water resources to supply the whole region. In this context, water scarcity refers to a situation where the volumes of water available cannot meet the demand for water. This can translate in restrictions or bans on water use reflecting the severity of droughts. For instance, the region experienced a drought of historic intensity in the summer of 2019, exacerbated by record summer temperatures reaching 42.6°C. Nearly half the region was subject to restrictions on the use of water for agricultural purposes, and soils were badly affected. In 2022, during the countrywide drought, almost 20% of rivers in the Paris metropolitan area were considered to be in a state of drought crisis due to their low flow rates (DRIEAT, 2023[2]), leading to restrictions on navigation on canals (under which boats were grouped together) and restrictions on the use of water for watering gardens, sports facilities, vegetable gardens, washing cars and irrigation (Kumari et al., 2021[14]).
Climate variability can lead to three main types of droughts:
Meteorological drought corresponds to a precipitation deficit over a long period. A vulnerability study carried out by the City of Paris concluded that this type of drought is very unlikely, except in the worst-case scenario, which shows this type of drought becoming more common by the end of the century.
Soil drought or agricultural drought is characterised by a surface soil moisture deficit (between 1 m and 2 m deep) sufficient to affect plant growth. It depends on precipitation and evapotranspiration, and is therefore sensitive to humidity, air temperature, wind, and plant and soil type. It can, but does not always, occur following a meteorological drought.
Hydrological drought occurs when lakes, rivers or groundwater levels are abnormally low. It depends on precipitation and soil conditions, influencing run-off and infiltration. It generally occurs following periods of agricultural drought and can have more lasting effects.
Water scarcity reflects the risk that the water available will not be sufficient for systems to function, whether due to quantitative or qualitative issues. The risk may be temporary (e.g. a drought of varying duration) or permanent.
Source: (United Nations office for disaster risk reduction, 2021[15]); Météo France.
Increased temperatures induced by climate change may occur simultaneously with droughts and generate additional water scarcity risks. Climate change-induced reductions in flow rates limit the dilution capacity of water resources. When pollutants are present, this capacity to maintain the properties needed to achieve ecological balance affects the quality of surface water. Fertiliser run-off, for example, reduces the oxygen concentration in water resources through eutrophication, while crop protection products lead to levels of toxicity that damage biodiversity and jeopardise the production of drinking water. In the same way, industrial and domestic wastewater discharges (e.g. via sewer systems) contribute to the deterioration of surface water quality. Finally, the climate variability that is forecast in the region predicts severe flooding that could jeopardise the quality of drinking water resources (Aquavesc, SEDIF, Sénéo, Ville de Paris, 2020[16]).
In the Paris metropolitan area, climate change is expected to lead to drier soils particularly during summers. Climate projections for the region point to a downward trend in summer precipitation and an increase in evaporation and evapotranspiration due to rising temperatures (Figure 2.3). Greater climate variability is expected to lead to more intense but shorter periods of rainfall over the course of the year. Such patterns are less effective at recharging groundwater reserves and increasing soil moisture levels ahead of summers (Météo France, n.d.[17]). It is also estimated that evapotranspiration could increase by an average of 16% by 2050 and 23% by 2100 in the Seine–Normandy basin, compared with daily averages over the 1970-2000 period4 (Agence de l'Eau Seine Normandie, 2016[18]). These projections could increase the average duration of the dry soil period by two to four months compared with 1961-905 (Météo France, 2022[10]).
Note: Summer precipitation is defined as the cumulative precipitation from June to September. Annual potential evapotranspiration (PET) is defined as the cumulative daily potential evapotranspiration over the year. It is calculated using the FAO formula taking into account radiation, which is calculated based on the Hargreaves method.6 The figures show the difference between precipitation and evapotranspiration relative to the mean value of these indicators for the 1976-2005 period. The scenario chosen (Representative Concentration Pathway (RCP) 4.5) represents an intermediate climate change scenario of a global average surface temperature increase of between 1°C and 4.5°C (compared with pre-industrial levels) by 2100.
Source: Authors, based on DRIAS data (Ministère de la Transition Écologique, 2023[19]).
By 2050, climate change is expected to produce drought conditions thus far only observed in France’s Mediterranean regions. By 2050, soils in the Paris metropolitan area in summer could on average be as dry as those in the Mediterranean over the 1976-2005 period. Average soil moisture in 2100 could be similar to those observed during current extreme droughts (Figure 2.4).
Note: The drought indicator used is the Soil Wetness Index (SWI). The index, ranging from zero to one, represents the quantity of water contained in the soil. The closer the value is to zero, the drier the soil. The indicator represents the average SWI per department for the period, or the projection according to the RCP 4.5 intermediate climate scenario (warming between 1.0°C and 5.5°C by 2100).
Source: Authors, based on DRIAS data (Ministère de la Transition Écologique, 2023[19]).
As a result of longer drought periods, the region is likely to experience a significant drop in water flow rates and groundwater levels. Water flows are forecast to fall by between 10% and 30% by 2070-2100 compared with 1970-2005 averages (Agence de l'Eau Seine Normandie, 2016[18]). Some simulations predict a 10-fold increase in the annual number of days of hydrological drought in the Seine river basin (Boé et al., 2018[20]). By 2050, groundwater recharge is also expected to decline) (Ville de Paris, 2021[21]). It is therefore estimated that the average groundwater level in the river basin will fall to levels similar to those found in the 10% of driest years to date compared with 1970-2005 (Agence de l'Eau Seine Normandie, 2016[18]). The region could then experience more droughts that were particularly intense and long-lasting. This would make a drought similar to that of 1921 likely by 2050 (Figure 2.5).
Note: Each point represents an annual drought simulation for a given climate model. The graphs on the left simulate potential droughts based on past weather data, while the graphs on the right simulate future droughts using climate projections. Each color represents a different climate model, with the same model potentially producing varying results. Agricultural droughts, or soil droughts, are reflected by the soil moisture level measured by the Soil Wetness Index (SWI), as defined by (Vidal et al., 2010[22]). Hydrological droughts reflect the annual average flow rate of the Seine River. The y-axis shows drought severity, representing the relative difference between simulated soil moisture or flow levels and the average values observed during 1958-2005 (soil moisture) or 1961-1990 (flow). The number of drought days refers to how many days the drought index falls below a threshold, indicating low soil moisture or low Seine River flow based on the historical periods mentioned.
Source: (Boé et al., 2018[20]).
Climate change also poses a risk to the quality of water resources. Falling water volumes, combined with increasing average and extreme temperatures, are expected to lead to an average increase in river temperatures of up to 2°C by 2100 compared with 1976-2005 (Agence de l'Eau Seine Normandie, 2016[18]). By 2050, water temperatures are forecast to rise by 0.2°C upstream of the basin and 0.9°C downstream (EPTB Seine Grands Lacs, 2022[9]) with impacts on aquatic ecosystems. A 10% to 30% drop in low water flow rates would reduce the pollutant dilution capacity of water resources.
The Paris metropolitan area is home to France's administrative, economic and cultural capital. The region comprises 1 276 municipalities, including the French capital, where all national administrations are headquartered. The region has a population of 12.2 million people, 19% of the French population (INSEE, 2023[23]). It accounts for 23% (INSEE, 2021[24]) of jobs in France and contributed 31% of French GDP in 2020 (INSEE, 2022[25]). The region's economy is based on a highly developed tertiary sector (86.3% of the region's value added and 78.8% of the national tertiary sector's value added, including commercial activities and administrative services). Industry generates 8.4% of regional value added and 13.7% of industrial value added in France as a whole. Finally, although agriculture represents only 0.1% of regional value added (1.9% of agricultural value added in France), agricultural land covers 50% of the territory (Institut Paris Région; INSEE; CCI, 2021[26]). This economic profile is reflected in the region's trade, making it both France's leading exporter (25% of national production exported internationally) and importer (29%). Agricultural production in the region, excluding agri-food products, makes up 3.4% of French exports (France, 2023[27]). Trade is further supported by the region's transport infrastructure, including its international airports, and rail, river and port networks, which make it an international and European hub. It is part of Europe's second-largest river network. Finally, the region is the world's leading tourist destination.
Water resources are essential to the region's activity. Surface water, in which 75% of the region's withdrawals are made (average for the period 2017-2021) is used for the production and distribution of drinking water (63%), energy production with restitution (16%), or industry (11%). In addition, the good health of ecosystems and a large part of the region's agricultural production depend on soil moisture, which tends to decrease, as well as the level of underground water resources, which are also subject to withdrawals (e.g. irrigation, drinking water, industry).
Source: Authors.
While the region has seen its water consumption decrease over the last twenty years, it could increase again. Demand for drinking water decreased by 8% on average between 1998 and 2008, with significant disparities between the city of Paris (-32%) and its suburbs, which saw an increase in consumption due to urban sprawl (Agreste, 2011[28]). This dynamic is tending to stabilise (Cour des Comptes, 2018[29]). The decline of large steel and chemical industries (Chevrot et al., 2018[30]) has led to an average decrease of 14% in water withdrawals for industrial use in the region almost every year since 2015 (Figure 2.7). However, population growth, estimated at 0.05% per year until 2070 (INSEE, 2022[31]), and the effects of climate change, such as urban heat islands, are expected to lead to an increase in water demand. By 2030 to 2050, climate change alone could increase water consumption by 2%, particularly in cities with little vegetation (Ville de Paris, 2021[21]). For comparison, water consumption could increase by 9 to 15% for the city of Naples in Italy for the same reasons (Fiorillo et al., 2021[32]). Finally, withdrawals for irrigation doubled between 2012 and 2020, due to the increase in irrigated areas (7%) (Agreste, 2022[33]) and the growing demand from plants due to heat waves, a trend that is expected to continue. These withdrawals are expected to continue to grow, partly to address climate challenges that are all the greater as heat waves will increase the demand for water from plants.
Source: Authors, based on data from the national database of quantitative water abstractions. Consumption (million m3) is calculated on the assumption that 80% of the water withdrawn for drinking water uses and 90% of that withdrawn for industrial uses is returned to the environment after use, while 100% of the water withdrawn for irrigation is understood to be consumed.
The region is the most densely populated in France, putting a significant strain on water resources for drinking water production. With 1 021 inhabitants per km2, the region’s population is ten times more densely populated than the rest of France (INSEE, 2023[34]). Both surface water (65%) and underground resources (35%) are used to produce drinking water in the interconnected zone.7 (Sénéo; Aquavesc; SEDIF; Ville de Paris, 2019[35]). In case of drought, the level of resources available could create supply difficulties for production facilities and an increase in water treatment costs due to higher concentrations of pollutants or the increase in pathogens in warmer water. In the past, despite the imposition of abstraction limits in 2003, 2011 and 2012,8 drinking water production continued at sufficient levels to serve the population. Domestic consumption in the region currently averages 123 litres per person per day (SEDIF, n.d.[36]), 9,10 i.e. below the national average of 148 litres (France, 2023[37]) and relatively low compared with other OECD countries (Figure 2.8). Reducing drinking water consumption would therefore likely affect other essential activities, with potentially severe negative effects.
Note: These figures need to be analysed with caution, given the metering methodology, the inclusion of business and office use (which is sometimes difficult to isolate and varies from city to city), and the inclusion of leakage in the calculation.
Source: (Barcelona Metropolis, 2023[38]); (Mayor of London, 2022[39]) ; (Waternet, 2021[40]); (Eau de Paris, n.d.[41]); (Danva, 2022[42]); (EPAL, n.d.[43]) (Victoria State Government, n.d.[44]); (Toronto, n.d.[45]); Data from the Pacific Institute (California Urban Water Use Data); (Climate ADAPT, 2018[46]); (OECD, 2022[47]).
The services sector relies heavily on the supply of drinking water. The region is home to the world's largest fresh produce market (Rungis market). If there were water supply disruptions, the market may face economic losses due to the need to limit sales to comply with health regulations. The region also regularly hosts conferences, trade fairs and sporting events. Tourism also requires large quantities of water in the summer season.
The manufacturing sector relies on water supply as a production factor. The automotive, electronics and aeronautics industries, along with the food industry, are among the largest sectors (Institut d’aménagement et d’urbanisme, 2016[48]) in the region. To a lesser extent, the pharmaceutical and chemical industries depend on surface water (EPTB Seine Grands Lacs, 2022[9]). Until now, the manufacturing sector has not been impacted by droughts. However, the water used by industries could lead to future conflicts, potentially affecting the amount available for production. A notable example is the dispute between the town of Grigny and the Coca-Cola company,11 which has private wells and is accused of depleting groundwater reserves. More broadly, industrial water use may face restrictions to prioritise access to drinking water for essential needs.
The region's energy production, which relies on water as a heat transfer fluid. Although the region produces only 13% of the energy it consumes (ROSE, 2023[49]), three of the five largest water users in the region have been power plants (BNPE, 2022[50]). Electricity generation by incinerating household waste, which accounts for 23% of the electricity produced in the region (DRIEAT Ile-de-France, 2022[51]) (RTE, 2021[52]), relies on surface water, coming mainly from the Seine river, to produce steam and electricity. This steam feeds the Paris heating network, which supplies all Paris hospitals with hot water, as well as 40% of the services sector buildings and 45% of social housing (CPCU, 2019[53]). Even a brief disruption to heat production due to excessively hot water from the Seine would affect the network's pressure profile, putting the system out of service for several months. The region also has cooling networks, with around 70% of their capacity relying on water from the Seine as a heat transfer fluid. These cooling networks meet 17% of the cooling requirements of museums, hotels and offices (APUR, 2019[54]) and are essential to the health of the most vulnerable populations (e.g. excess mortality during heatwaves, health risks) and the continuity of the region's health services.
Occupying 50% of land in the region, agricultural production has historically been the activity most at risk of water scarcity. The Regional Plan for Local, Sustainable and Inclusive Food makes food sovereignty a priority and supports the economic development of the sector (Région Ile-de-France, 2021[55]). Today, 22% of the region’s agricultural production is distributed through short supply chains.12 Furthermore, the region’s agricultural production supplies the agri-food industry, which accounted for an average of 1% of the region's value added (INSEE, 2023[56]) over the 2015-20 period and 4% of the region's exports (Ministère des Finances, 2023[57]). The region’s farms, particularly cereal and oilseed farms – which account for around half of the region's agricultural production value – primarily depend on precipitation to ensure sufficient soil moisture levels for agricultural production. They also rely on irrigation, which is currently essential for a share of fruit, vegetable and corn production, but can be restricted or even temporarily banned by decree if there is a severe drought.
Indispensable for the region's agricultural, energy and industrial production, and for the survival of its population, water resources are also essential to the transport of goods and tourists in the region. The Seine transports 25 million tonnes of goods in the Paris metropolitan area (Institut Paris Region, 2021[58]), i.e. more than half of all goods transported via inland waterways in France (Voies Navigables de France, 2022[59]). People also enjoy travelling around Paris by boat. The Port of Paris, for example, is the world's leading tourist port, transporting nearly 8 million passengers in the capital (Haropa Port, 2023[60]). Very low water levels reduce the flow and depth of watercourses, potentially limiting ship loading and even traffic, which leads to delivery delays and the cancellation of cruises. In addition to representing a loss for the river sector, these stoppages impact the logistics chain and might affect other economic activities or trigger a modal shift to lorry or bus transport, resulting in increased traffic and pollution, including CO2 emissions. Although commercial navigation on the Seine was relatively unaffected by past droughts, pleasure boating on canals – human-made structures that depend entirely on water being pumped from the region's main rivers – is particularly at risk when water levels are low because drought orders place restrictions on their replenishment. In 2022, canal transport was restricted in specific cases in an effort to relieve pressure elsewhere.
Droughts pose a risk to buildings. Dry soils contribute to the phenomenon of clay shrinkage and swelling. This phenomenon, caused by changes in soil moisture damages residential houses. In France, the Paris metropolitan area is the region most exposed to this risk: 76% of its territory is located in medium to high-risk zones (Figure 2.9). This resulted in average payouts of EUR 62 million per year between 1995 and 2016 (Institut Paris Région, 2022[61]) (Ministère de la Transition Ecologique et de la Cohésion des territoires, 2021[62]).
Source: (Institut Paris Région, 2022[61]).
Finally, urban green spaces and ecosystems, which contribute to the region's resilience, are also vulnerable to droughts. Park and garden watering, which is essential for their upkeep, is one of the first uses to be restricted if there is a severe drought. Nevertheless, the adaptation plans or climate change resilience strategies drawn up by the region and some municipalities do identify renaturing urban spaces as a strategy (Région Ile-de-France, 2022[63]) (Ville de Paris, 2019[64]). Green spaces help water infiltrate the soil, preventing it from overwhelming the drainage network during heavy rainfall (a cause of flooding and increased pollution of watercourses) and reducing the amount of water requiring treatment. This infiltration also allows the soil to filter the water, thereby reducing the need for treatment. Green spaces also help combat the urban heat island effect. Providing a natural cooling effect in the summer months, trees help curb the increase in drinking water consumption observed in other cities (Fiorillo et al., 2021[32]). Soil drought also poses a risk to the region's ecosystems such as wetlands, which help to maintain biodiversity and store water. Wetlands make up more than 2.8% of the region (Institut Paris Région; INSEE; CCI, 2021[26]). Finally, droughts make tree species more vulnerable to the spread of pathogens (Institut Paris Région; INSEE; CCI, 2021[26]) and can cause trees to stop growing or even die. Droughts therefore lead to trees being felled early, known as "sanitation felling", which reduces the forest's productive capacity and the value of the wood produced.
Estimating the future economic costs of droughts helps identify the sectors most impacted by climate change and the potential ripple effects of such events. The OECD developed socio-economic scenarios to project water demand and builds on existing climate scenarios to anticipate water supply. This approach assesses the direct effects of droughts on the region's economic activities and models their broader impact on the French and European economies.
The Paris metropolitan area: Looking towards 2050
The socio-economic development of the Paris metropolitan area increases drought risks. Although there are many very different possible development scenarios for the region looking to 2050 (Box 2.2), extrapolations based on current sectoral trends and existing planning documents suggest the region's dependence on water resources will increase by the middle of the 21st century. Irrespective of potential gains in resource efficiency, the economic and demographic development of the region is expected to lead to an increase in water requirements.
The population of the region is set to increase by around 5% by 2050 and with it the need for drinking water and wastewater treatment (INSEE, 2022[65]). This potential rise in demand will be exacerbated by climate change events, particularly heatwaves, which will increase demand for drinking water for hydration and cooling. Global warming is expected to increase overall water consumption by 2% between 2030 and 2050 (Ville de Paris, 2021[21]). Moreover, population growth in the river basin will increase the need for water treatment and the discharge of wastewater by local authorities (Agence de l'Eau Seine-Normandie, 2019[66]), which could lead to a decline in the quality of surface water.
Although withdrawals for irrigation are currently limited, irrigation needs could increase significantly by 2050. The increase in vegetable and fruit production to meet growing demand for short supply chains, combined with a rise in agricultural droughts, will drive the expansion of irrigation. The area of irrigated land in the region increased by 14% between 2010 and 2020. If this trend continues, the irrigated area of the region will increase by 45% by 2050 (compared with 2020).13
The development of cooling systems that use water from the Seine for air-conditioning could be affected by the water temperature. Climate change will increase the need for air-conditioning, as well as the need to cool production facilities. Indeed, the master plan for the City of Paris’ cooling network estimates that 70% of production will depend on water from the Seine by 2050 and forecasts that the network will deliver 1 000 GWh/year of power by 2050, i.e. an almost 2.5-fold increase on current capacity by 2050 (Ville de Paris, 2019[67]).
River transport is also expected to increase sharply by 2050. The current rate of growth in river freight – 3% between 2020 and 2021 (Voies Navigabels de France, 2022[68]) – suggests that the current volume of goods transported will increase 2.4-fold by 2050. This trend is in line with the ambition to further develop river transport in the region, as demonstrated by the Seine-Escault project, which aims to connect Paris and Le Havre to the major seaports of the North Sea, via the Seine, Oise, Escault and Lys rivers. Similarly, river tourism is currently growing by 4% per year, which will result in a three-fold increase in traffic by 2050.
A number of possible scenarios have been produced to show how the region will change by 2050. There is no consensus on any of them and it is difficult to choose any one as the reference scenario for study, not least because they are based on different objectives and assumptions. For example, the French Agency for Ecological Transition (ADEME) has developed four scenarios for France with a 2050 outlook,1 which can be applied at the regional level. These four scenarios correspond to different social choices (economic, technical, etc.) regarding how to achieve carbon neutrality by 2050. Some are based on more restrained consumption and use, while others position technological development and the rise of digital technology as the driving force behind more environmentally sustainable development. The Momentum Institute provides a description of a post-collapse scenario for the region. This breakaway scenario, called "Bio Region 2050", proposes drastic systemic change. The region would be divided into small autonomous regions and cars would have disappeared. Calls for projects such as Vers Paris 2050, affronter ensemble les défis de notre siècle [Towards Paris 2050: Facing the challenges of our century together] also offer the opportunity to develop a narrative about the future of the French capital by 2050.
A number of sectoral and land-use plans have also undertaken this prospective exercise. In the region, the Environmental Master Plan for the Region (SDRIF-E in French), is the regulatory reference document for urban and regional planning. The SDRIF-E, to be voted on in the summer of 2024, will provide a framework for regional development until 2040. Similarly, the Regional Plan for Local, Sustainable and Inclusive Food brings together 66 measures to improve the Paris metropolitan area's food autonomy by 2030.
Finally, the Explore 2070 project aims to study the impact of climate change on water resources and aquatic environments in France, looking ahead to 2070. It has carried out a prospective exercise to quantify future water requirements on a national scale. To do this, the authors developed two socio-economic and demographic trend scenarios, based on an extrapolation of past trends. One is a denser housing scenario and the other an urban sprawl scenario. These scenarios differ in terms of transport needs and how much land will be built-up or available for agriculture.
Forecasts for manufacturing in the region are less clear. The trend towards industrial site closures in recent years could be offset by reindustrialisation policies. The hypothesis that the manufacturing sector will be identical to its present form (in composition and value) seems appropriate.
It is also difficult to predict how ecosystems and urban green spaces will develop in the context of climate change. Consequently, although the City of Paris and the region have made tree planting a major lever of their resilience strategy, it may be difficult to foresee how this will affect demand for water. Outcomes will depend on the species planted and the cooling effects provided by the green spaces created. Similarly, while the regional biodiversity strategy for 2020-30 (Conseil Régional d’Ile de France, 2019[72]) and the national objective of net-zero increase in built-up land suggest that environments will be relatively protected, it is difficult to assess how these environments will adapt to the climate of the future. The impact analysis below is based on the assumption that natural and urban ecosystems will be identical to those that exist today, due to the lack of information on the impacts of regional development and climate change.
While the water supply is expected to decrease during severe summer droughts, water demand is expected to rise, increasing the risk of water scarcity. To assess the region's future dependence on water resources, the economic analysis below uses a socio-economic scenario that extrapolates current development trends to establish what the region will look like in 2050 (Table 2.2).
Extrapolating past socio-economic development trends in the Paris metropolitan area to 2050.
Source: Authors; scenario details available in Appendices.
The cost assessment is based on three plausible drought scenarios. The scenarios were developed to analyse the potential range and intensity of climate change impacts on water resources in the region by 2050 and 2100. Each scenario describes a specific hydrological and meteorological situation over the course of a year. The analysis is based on deterministic drought scenarios, given that its purpose is not to describe the distribution of possible impacts or their probability, but to assess what might happen in several decades' time in the event of periods of extreme drought. These scenarios are defined by climate model projections for the 2050/2100-time horizons (Box 2.3).
The study applies current drought management rules and legislation to establish the regulatory impacts of water scarcity. The study is not intended to discuss the suitability of different adaptation or development scenarios. It therefore applies the current rules of the departmental drought orders and those applicable to installations classified for environmental protection, as well as the regulations on discharge temperatures applicable to industrial activities, without making any judgement regarding their (current or future) suitability. This subject is addressed in the following chapters of this report. The crisis threshold is breached for 55 days and 92 days between July and December for the Marne and Seine rivers respectively in scenarios 1 and 2, and for 140 days and 166 days between April and December in scenario 3. Similarly, the Champigny aquifer, the reference for all the region’s aquifers in our scenarios, is in a crisis situation from mid-April to the end of December in scenarios 1 and 2, and throughout the year in scenario 3. According to departmental drought orders, when the crisis threshold is breached, irrigation is banned (if using surface water) or reduced by 40% (for groundwater), navigation on canals is halted and industrial activity is reduced by 25%. Surface water temperatures remain below the critical threshold of 25°C in scenario 1 but exceed it on 71 days from June to September (including two days above 29°C) in scenario 2. Similarly, temperatures exceed 25°C for 107 days, including 28 days over 29°C, in scenario 3. These high temperatures lead to a 20% fall in cooling and electricity output (from burning household waste) between 25°C and 26°C, and a further 20% for each additional degree between 26°C and 29°C. Above 29°C, energy production comes to a complete halt.
To assess the impacts of climate change-induced droughts, three scenarios have been developed based on existing drought scenarios for the basin, as well as on climate projections of temperatures. They are not intended to reflect the consequences of different climate change scenario pathways (Representative Concentration Pathways (RCP) or Shared Socio-economic Pathways (SSPs)), as these scenarios are statistically very similar for a 2050 outlook. Rather, they aim to reflect the various events that could lead to resource scarcity by 2050 and 2100.
The drought scenarios include quality and quantity parameters corresponding to the main factors of scarcity for the different activities exposed to this risk in the region. As the table below shows, each scenario describes the annual change in river flow rates and replenishment of EPTB reservoirs (low-water support), groundwater levels, soil dryness and surface water temperature. The change in each of these variables is simulated for a limited number of stations: the Seine at Alfortville, the Marne at Gournay, the Oise at Creil and the Champigny aquifer at Montereau-sur-le-Jard. These stations are the drought order reference stations for the most-used water resources, identified using data from the national database of quantitative water abstractions.
Source: (Boé et al., 2018[20]). A detailed explanation of how these scenarios were established can be found in the Annex B.
Quantifying the economic impact of climate change-related events
This prospective study assesses the direct and indirect economic costs of droughts.
Direct costs represent all the impacts, both negative and positive, experienced by the actors and regions directly affected by events. Direct costs can be tangible and intangible. The tangible costs include capital losses due to the destruction of assets, and production losses due to the unavailability of labour, means of production, or resources. In the case of droughts, direct costs mainly relate to production losses due to water being unavailable.
The first stage of the impact study involved identifying the activities (or systems) in the Paris metropolitan area that depend on water resources, and the mechanisms that might lead to water scarcity for these activities. The next stage involved defining the relevant indicators (e.g. temperature, pollution, flow) to study the direct impact of droughts on each activity and explain the relationship between the level of these indicators and the various activities. For example, energy production is affected once a certain temperature and flow rate are reached. Below these levels, activities are not impacted. The aim was therefore to obtain, for each activity, a list of indicators, the level of these indicators at which the activity is impacted, and the relationship between changes in these indicators and the impact on the activity.
The information needed for this study to calculate the direct costs of droughts was gathered from researchers, economic actors and institutional representatives in the region, mainly on the basis of interviews, with a view to understanding the region's specific hydrological, environmental and socio-economic characteristics. The OECD asked around 100 actors (see Appendices) about their activities, their dependence on water resources, the resources they currently have at their disposal to manage this risk, and the impact that reaching critical levels of resource quality or quantity would have on their business. These interviews enabled us to identify the scarcity factors likely to affect each activity, the thresholds above which activities could be impacted and the impact of scarcity on the activity.
The limited scope of the study and the proximity to all the local actors from whom data was collected had a number of advantages over using existing literature. Firstly, detailed knowledge of each sector's dependence on water resources enabled us to accurately identify direct impacts and in particular to avoid using generic damage functions. In addition, we were able to gather data on the origin of the water used by the various activities (place of abstraction and source), enabling us to distinguish between uses reliant on the region's different aquifers and watercourses. This allowed us to examine in greater detail how each activity will be impacted, depending on the condition of the water source on which it depends. Furthermore, while most existing studies limit the definition of scarcity to the quantity of water available, we integrated water quality issues affected by river temperature, and therefore studied the concomitant effect of heatwaves and droughts. Finally, the study takes into account existing infrastructure, such as the diversion of the Seine into reaches and reservoirs, which have a major influence on water availability and, in turn, the region's vulnerability.
In a second step, the direct costs were applied to a macroeconomic model to assess the indirect or cascading impacts on the entire regional economy and calculate the total cost of each of these events for the region as a whole. Indirect costs are the consequences of how direct costs affect other economic activities and regions not directly impacted. The full, or macroeconomic, cost of an extreme weather event is therefore not limited to its direct cost, but includes its knock-on effect on the entire regional, national or even global economy, via value chains and the interdependence of different economic sectors (Box 2.4). Finally, modelling indirect costs helps identify any positive effects of droughts, given that some sectors or regions may benefit from a compensation effect. For example, the need to reconstruct buildings damaged by a climate event increases activity in the construction sector and therefore its revenue.
The macroeconomic Adaptive Regional Input-Output (ARIO) model was used to model the indirect impacts and calculate the macroeconomic cost of each scarcity event. The model (Box 2.4) represents the economy as a set of economic sectors and regions. The EUREGIO database, based on national and regional accounts data, was used to model the links between the various economic sectors within the region, and between the French regions and those of 23 European countries.14 The month-by-month values of the sectoral direct costs described above were used in this model to represent production interruptions, the effects of which spread along value chains.
Macroeconomic models are needed to identify the indirect costs of extreme weather events. Three groups exist: input-output (I-O) models, partial or computable general equilibrium (CGE) models or. agent-based modelling. I-O and CGE models are both based on social accounting matrices, representing monetary flows between all sectors of an economy. However, they use different frameworks to account for the behaviour of economic agents (consumers and companies).
I-O models assume fixed-proportion, time-invariant output functions for all economic sectors and predict how damage in one sector will affect trade and related output in all other sectors. Technical coefficients are used to calculate the dependence of one sector on another. The model does not explicitly account for prices.
CGE models assume stable behaviour, reflected by stable demand and supply functions, and predict how the impacts of natural disasters will affect demand, supply and prices on different markets in equilibrium. Agent behaviour is assumed to be optimal. Sector dependencies are taken into account by means of constant elasticities.
Agent-based models are computer models in which different agents – households, businesses, farmers, governments – act, learn, interact and evolve with their environment (Ghaffarian et al., 2021). Their behaviour is more complex than the assumption of rationality allows.
The main difference between the I-O and CGE models lies in the possibility for substitution available to economic agents in the CGE models. Because they have the possibility to substitute inputs according to price variations, economic agents tend to be more resilient in CGE models. This explains why the higher ratios of indirect to direct disaster-related losses in I-O models are not observed in CGE models.
For this study the Adaptive Regional Input-Output (ARIO) model was selected. In ARIO, each economic sector manufactures a single generic product and draws its inputs from an inventory. Each sector meets a total demand made up of final demand (household consumption, public spending and private investment) from all regions (local demand and exports) and intermediate demand (inventory replenishment). An initial equilibrium state of the economy is built based on multi-regional input-output tables (MRIO tables). Two types of shock can be used: either at production level (an industry is forced to produce less by external factors) or at capital level (an industry loses some of its production factor due to external factors and is therefore forced to produce less and rebuild its capital stock). The model then describes how external shocks spread through the economy at each time step (one time step corresponds to one day). The total economic impact can be measured in two ways: in terms of (i) unsatisfied final demand or (ii) relative production loss.
The ARIO model has been widely used in the literature to simulate the propagation of negative shocks throughout the economy. It was originally designed to assess the total economic cost of Hurricane Katrina in 2008. Since then, ARIO has been used to estimate the total economic cost of (i) the flood risk in Rotterdam (E. E. Koks et al., 2015), (ii) the extremely damaging wildfires of 2018 in California (Wang et al., 2020) and even (iii) the supply chain effects of global COVID-19 lockdown measures (Guan et al., 2020).
2.5.1. Results of the direct and indirect economic costs of future climate change induced droughts
The direct cost of an extreme drought event in the Paris metropolitan area amount to EUR 966 million for scenario 1, EUR 990 million for scenario 2 and almost EUR 1.5 billion for scenario 3 (Table 2.3). The direct cost is mainly a function of production losses in the manufacturing sector, which represent between 48% (scenarios 1 and 2) and 57% (scenario 3) of the total value added lost for the region in each of the three scenarios, agriculture (15-16%) and clay shrinkage and swelling (between 36% and 24%). These costs represent up to 0.2% of the GDP of the Paris metropolitan area (2020). The breakdown of direct costs by sector is mainly based on the relative economic size of each sector in the region’s economy. The annual costs incurred by the manufacturing sector represent between 1% and 2% of the sector's annual value added, compared with up to 18% for the agricultural sector. However, the contribution of manufacturing to the GDP of the region is over 80 times that of agriculture.
Loss of value added or destruction of capital for each of the three scarcity scenarios in the Paris metropolitan area
Note: For all sectors, with the exception of detached homes, the costs represent a loss of value added. For detached homes, the costs represent a loss of asset value. Costs are expressed in 2020 euros.
Source: Authors.
The direct costs in each of the three drought scenarios are mostly production costs, not capital costs, and are mainly a result of regulatory measures. Restrictions on water use due to the enactment of drought orders and discharge temperature limits account for over 60% of the direct cost in scenarios 1 and 2. Table 2.3 shows the scarcity factors that lead to direct costs for each sector. The costs to industry, water transport and a share of agriculture (irrigation) are due to the restrictions on water use (called drought orders) that are applied when the different severity levels under drought orders are crossed. Only the costs related to soil drought affecting non-irrigated crops and the damage caused by the effect of clay shrinkage and swelling on buildings are of natural origin, i.e. not influenced by restrictions on water use.
The direct costs (loss of value added) for the manufacturing sector in the region are estimated at between EUR 466 million and EUR 827 million (Table 2.3). Water abstraction for industrial use is limited by the decree governing installations classified for environmental protection,15 which provides for a reduction in abstraction of 5%, 10% and 25% for each level established by the drought orders. In the absence of detailed information on water management at each industrial site, and assuming that industrial processes make optimal use of water, industries are obliged to reduce their production in proportion to the abstraction reductions required by the decrees.
The costs for the agricultural sector stem from yield losses and represent losses of up to 14% of the sector's value added in scenarios 1 and 2, and 18% in scenario 3. Yield losses affecting arable crops are mainly due to soil drought (agricultural drought), particularly for non-irrigated crops. Lower yields due to agricultural drought are responsible for over 80% of losses affecting arable crops in scenarios 1 and 2. In addition to yield losses due to dry soils, arable crops and above all vegetable and fruit crops suffer yield losses due to restrictions on irrigation. Drought orders provide for a ban on irrigation when the crisis threshold is exceeded. Each week below the crisis threshold prevents one round of irrigation, corresponding to a water volume equivalent to 30 mm per hectare. Irrigation restrictions account for value-added losses of over EUR 70 million in scenarios 1 and 2, and almost EUR 140 million in scenario 3. Irrigation restrictions are responsible for all losses affecting vegetable and fruit crops, accounting for 48% of losses in the agricultural sector as a whole in scenarios 1 and 2 (and 63% in scenario 3).
Electricity production using household waste and cold production using water from the Seine account for most of the impact on energy production in the region. In both processes, the Seine is used as a heat transfer fluid subject to the decree of 24 August 2017,16 which limits the discharge temperature of water into the environment to 30°C. When the temperature of the Seine rises above 25°C, energy production drops by 20% for each additional degree, to zero above 29°C. The temperature of the Seine limits energy production for 70 days in scenario 2 (and 107 days in scenario 3), leading to the shutdown of production for a few days in scenario 2 and for one month in scenario 3. Cold production losses represent 16% (34%) of annual production in scenario 2 (scenario 3) and 9% (19%) of energy recovery. Other electricity generation methods used in the region are not dependent on water resources (renewable solar and wind power, air cooling, etc.).
For water transport, only the supply of water to canals is subject to drought orders and therefore only pleasure boating is impacted. Water transport of goods and tourists is not affected because navigation on the main routes in the region, in particular the Seine, is guaranteed by the reach infrastructure and reservoirs. Similarly, the characteristics of the Seine-Scheldt canal (watertightness, conservation of water through locks and a support lake) mean it is resilient in each of the three scarcity scenarios (see Appendix 5.5 for details of resilience calculations for this sector). Consequently, the cost of scarcity does not exceed 1% of the sector's annual value added.
Drinking water production is not impacted in our study, because the drought orders do not impose any restrictions on it. As a result, only technical limitations affecting extraction (pump depth, for example) or water treatment are expected to give rise to costs. The groundwater or river levels reached do not pose a physical risk to water abstraction, and any costs incurred as a result of having to treat warmer water are negligible (according to experts).
Finally, although difficult to estimate, costs resulting from damage to buildings due to clay shrinkage and swelling could be at least as high as those recorded during the 2022 drought, amounting to EUR 346 million for the region.
The total cost of droughts in the region, i.e. the sum of direct and indirect costs, is estimated EUR 1.4 billion in scenarios 1 and 2, and almost EUR 2.5 billion in scenario 3 (Table 2.4). These costs represent up to 0.34% of GDP of the region (2020). For each scenario, the indirect costs are similar to the direct costs. Considering the propagation of costs therefore doubles the total cost of the different events compared with the direct costs alone. These costs are of the same order of magnitude as those estimated by similar studies (Box 2.5).
Loss of value added, EUR million (2020 euros)
Note: Costs are expressed as loss of value added. Total costs represent the sum of the direct costs (see Table 2.3) and the indirect costs calculated using the ARIO model. Cost amplification is defined as the ratio between the total cost for France and the total direct costs for the Paris metropolitan area. Amplification is used to assess the impact of indirect costs on the total cost of the event.
Source: Authors.
A wide range of methods can be used to measure the cost of droughts (Logar and van den Bergh, 2012[73]). The following table summarises the findings and methods used by various studies measuring these costs.
Breakdown of indirect costs by sector
Water scarcity has an indirect impact on all sectors of the economy. Although many sectors – such as transport, tourism, distributive trade and construction – are not directly affected by water scarcity, they are all indirectly affected by interruptions to production. Figure 2.10 shows the loss of direct (blue bar) and indirect (grey bar) value added for different sectors in the region. The sum of the blue and grey bars represents the total cost incurred by the sector. Some sectors are affected almost exclusively by direct production interruptions, such as the 'Other manufacturing' sector, where the grey bar (indirect losses) is almost non-existent compared with the blue bar (direct losses). Conversely, indirect costs account, for example, for the entire economic impact borne by the distribution and real estate sectors, i.e. 5% and 2.5% of the annual revenue of these sectors respectively. In the simulations, losses are mainly propagated due to the decline in industrial production, which therefore reduces the volume of intermediate inputs purchased from suppliers. Service sectors that supply the sectors unable to produce are particularly affected by indirect impacts. There is no propagation from producers to buyers, as production interruptions are offset by the existence of reserves.
Percentage of value added lost by sector (600-day simulations) for scenario 2.
Note: The total value added lost represents the sum of direct and indirect costs.
Source: Authors.
The inclusion of indirect costs changes the relative weight of the different sectors and increases the share of costs borne by the manufacturing sector. Indirect costs mainly come from the manufacturing sector, and in particular the agri-food sector, where changes in production lead to indirect costs more than four times higher than the direct costs incurred by the sector (see Table 2.5). In contrast, the indirect costs of disrupting water transport are considerably lower (37% in scenario 1) than the direct costs.
Direct and indirect costs generated by each sector directly impacted by droughts in the region (scenario 1)
Note: The "total costs" per sector are defined as the sum of the direct costs borne by that sector and the indirect costs to the region of the disruption to activity in that sector. "Amplification" is defined as the ratio between the total costs and direct costs. The sectors are classified using the EUREGIO nomenclature, on which the ARIO model is based. The sum of the total sectoral costs differs from the total cost presented in Table 2.4, because calculating the indirect cost borne by each sector involves simulating the impact of scarcity for each sector independently.
Source: Authors.
Spatial propagation of indirect costs
The impacts of scarcity are also felt outside the region, and their propagation to the national level generates additional costs of between EUR 102 million (scenarios 1 and 2) and EUR 172 million (scenario 3). As illustrated by Figure 2.11, the costs of these drought events for Corsica are equivalent to 10% of the impact observed in Paris metropolitan area, followed by Bourgogne-Franche-Comté (9%), Nouvelle-Aquitaine (9%), Grand Est (7%) and Centre-Val de Loire (7%). Europe, and in particular countries neighbouring France, are likewise impacted by scarcity events. The costs of scarcity to other European countries ranges from EUR 102 million (scenario 1) to EUR 172 million (scenario 3) and are mainly borne by Germany, Spain, the United Kingdom, Belgium and the Netherlands.
Propagation of drought costs by European region for scenario 1
Note: Scenario 2 is characterised by: soil drought equivalent to 1976, river flow equivalent to 1921, a drop in groundwater levels equivalent to 2020-22 and a heatwave corresponding to the 90th percentile of the temperature distribution by 2050, based on simulations using a set of models in an RCP 8.5 scenario.
Source: Authors.
2.5.2. Limitations of the analysis and avenues for further research
The cost analysis presented in this report provides a better understanding of the scale of the challenges posed by climate change and water scarcity for the region. Nevertheless, it has a number of limitations and opens up many avenues for research that could improve cost evaluation and better inform government decision-making.
Uncertainty of climate and socio-economic projections
Analysing the impact of climate change on water resources to 2050 remains tricky given the many uncertainties surrounding climate projections and their use in hydrological models. Firstly, linking climate projections and hydrology is an exercise that requires data adapted to local climatic conditions. However, atmospheric circulation climate models are produced at a global level and translated to the regional level using techniques such as statistical downscaling (GIEC, 2022[81]). These techniques have their limitations, because they are unable to reconstruct important climate-related information at sufficiently fine scales. Irrespective of issues of scale, climate models give rise to a wide range of projections, depending on their design (Lehner et al., 2020[82]). While the study takes account of this variability by comparing the results of a set of models to determine how drought events will evolve, certain parameters remain particularly difficult to predict, such as evapotranspiration and changes in precipitation (GIEC, 2022[83]).
Looking ahead to 2050 also requires us to hypothesise how the socio-economic and environmental profiles of the Paris metropolitan area will evolve. There is a very wide variety of potential development trajectories for the region to 2050 and these trajectories have a significant influence on the region's exposure and vulnerability to the risk of water scarcity. Moreover, the study projects past development trajectories without taking into account the effect of climate change. The growing impacts of climate change are likely to influence development trajectories and/or lead to the implementation of adaptation strategies to mitigate their effects. Our study is a first step that could serve as a reference for complementary analyses assessing the effect of different development trajectories and adaptation strategies on the impact of water scarcity. This would then enable us to identify the changes that would best support the region's resilience.
Interdependence of climate phenomena
The economic study is based on a combination of climate events that are not independent. Indeed, the study constructs the various scarcity scenarios as aggregations of isolated results from different models, each describing the change in a hydrological or meteorological variable (water temperature, river flow rate, groundwater level, etc.). However, as described in this chapter, aquifers and rivers are interconnected, and their levels are therefore dependent on one another. Similarly, water temperature is influenced by air temperature and flow rates, and air temperature is itself a factor in drought, particularly in agriculture. Accounting for all these factors within a single model would enable the production of more robust and coherent scenarios.
Furthermore, the study is limited to the impact of drought on economic activities, ignoring how the factors behind scarcity can themselves impact these activities. For example, air temperature affects water temperature, and therefore the ability to produce energy, but high temperatures are also one of the main factors driving higher energy demand in summer, whether for cooling industrial facilities, offices or homes. Similarly, this analysis does not consider the influence of these scarcity events on other risks, such as forest fires or industrial accidents.
Finally, the analysis focuses on the impact of drought events affecting the Paris metropolitan area, independently of conditions simultaneously affecting the rest of France or even the world. However, it is highly likely that if the region is affected by extreme temperatures or droughts, other French regions, particularly in the south of the country, will also be affected. The study does not therefore take account of the spatial correlation of weather and climate conditions. Studying the region in isolation could mask much greater impacts. Indeed, the region is not self-sufficient and imports many goods and services. For example, it imports 95% of the energy it consumes (DRIEAT Ile-de-France, 2021[84]).
Moreover, the study does not address the macroeconomic effects of scarcity, notably the effect of lower production on prices. A drought in France could, for example, reduce the production capacity of nuclear power plants and hydroelectric dams, leading to a spike in energy prices. Similarly, a Europe-wide agricultural drought could lead to a sharp increase in the price of agricultural commodities. In addition to the reduced availability of resources constraining economic activity, rising prices for goods, energy and foodstuffs could generate additional costs for companies in the region, as well as households, which are excluded from this analysis.
Limitations of the ARIO model for regional economic impact analysis
The ARIO model has two main intrinsic limitations: the simplicity of how the economic mechanisms at work are represented and the amount of data required for the model to function. While ARIO integrates certain characteristics of the real economy better than other models, such as production inventories or the demand for reconstruction, the model only takes into account simple mechanisms. It does not incorporate possible price changes, as computable general equilibrium models do, and it uses rigid rules (which are similar in all industries) to establish how companies will behave. Based on the principles of input-output models, ARIO is designed to study the consequences of short-term economic shocks. As a result, agents are limited in their substitution possibilities, both as suppliers and as buyers. The second limitation is the large amount of input data required to run the model. In general, data to assess a wide range of parameters (e.g. characteristic times or the overproduction factor α) are not available at the local scale and standard values from the literature must be used, which are often identical for every sector. The two limitations are linked: the more complex the model, the more data you need to calibrate it. ARIO was therefore chosen as a compromise between these two concerns.
The use of the ARIO model is also subject to several limitations specific to this analysis. Firstly, the decision to conduct a study at the regional level necessitates the use of EUREGIO, an input-output table with a limited number of economic sectors. This crude description of input-output links limits the identification of blockages in the supply chain, but there is currently no better alternative at the regional level. Furthermore, this table reflects the structure of the global economy in 2010, whereas the analysis looks ahead to 2050 (scenarios 1 and 2) and 2100 (scenario 3). Nevertheless, the analysis does provide an approximation of future risk levels, assuming the economy will continue to develop following historic trends to 2050. Finally, the ARIO model, which is better tailored to the industrial sector, may not fully take into account the characteristics of the agricultural sector: for example, crop losses lead to a reduction in input purchases upstream in the model, which does not reflect the reality. However, this criticism does not fundamentally call into question our findings, because impacts on agriculture are minor compared with impacts on manufacturing in terms of the direct costs.
Beyond impacts on economic activity
The study focuses on the impact of scarcity on the region's economic activity, providing a partial picture of the extent of the real impacts of this phenomenon. As described in Section 2.4, preserving water resources is also essential to maintaining ecosystems and local biodiversity, as well as the physical and mental health of the Paris metropolitan area’s residents. The existence of urban nature and green spaces has a significant influence on their physical and mental health, particularly in urban areas (Barton and Rogerson, 2017[85]). Moreover, it is hard to put a price on the importance of the Seine – which flows through Paris and the surrounding region and goes far beyond its role as a means of transport, a water source or factor in energy production – and keeping it in good biological and aesthetic condition.
Analysing isolated extreme drought scenarios seems unsuited to assessing the impact of such events on ecosystems. The complexity of how ecosystems function and the existence of tipping points make it impossible to attribute ecosystem degradation to different levels of scarcity. Most of the impacts affecting ecosystems are the result of repeated episodes of scarcity or of progressive changes in weather patterns, rather than a single event. Trees, for example, are more sensitive to changes in temperature and average drought conditions, which affect their growth and ability to survive in an environment, than to a major drought event, which they often have the capacity to withstand.
Moreover, assigning a monetary value to the cultural and natural heritage of the Paris metropolitan area is extremely complex. When goods and services have no market value, as is the case for nature, quantifying the value of ecosystem services to the regional economy and its inhabitants relies on multifactorial and qualitative analyses, such as the willingness of different stakeholders to pay for services. Many studies seek to establish the value of different biomes. (Brander et al., 2011[86]) presents a meta-analysis of 120 studies assessing the value of wetlands in temperate environments, estimating a value of USD 362/ha/year. Similarly, a metanalysis by (Woodward and Wui, 2001[87]) estimates that wetlands have an economic value of USD 393/ha in respect of flood control, USD 417/ha for water quality and USD 127/ha for water volume. The Ecosystem Services Valuation Database (ESVD)17 also enables us to estimate the average economic value of ecosystem services provided by lakes and rivers at USD 26 085/ha, forests at USD 21 647/ha and natural spaces in urban areas at USD 31 318/ha. However, we were not able to use these valuations in our study because they are specific to a type of service (preservation of water quality, water volume, local biodiversity, etc.), the economic benefits of which cannot simply be added together. Moreover, the importance of the natural heritage of the region, such as the Seine and the region's many forests, to the attractiveness of the region and the well-being of its inhabitants is not included in these analyses. The valuation estimates are also highly variable. Based on a sample of 15 studies carried out in France, the General Commission for Sustainable Development (Bouscasse et al., 2011[88]) shows, for example, that the value of wetlands for water purification ranges from EUR 15/ha to EUR 11 300/ha.
Likewise, this analysis does not identify social impacts, in particular conflicts over the use of water resources during droughts. The tensions that emerge around resource allocation in times of shortage can also lead to economic losses (property damage, strikes, production stoppages pending a decision by the authorities). For example, Paris' water supply comes from sources outside the region and its abstraction can be a source of conflict with local farmers. The clashes that took place at Sainte-Soline or Sivens in France also represent a direct cost to society (mobilisation of law enforcement and health personnel, material damage and human injury, etc.). They are likely to influence public opinion and therefore citizens' expectations and use in relation to water resource management. Whether it leads to conflicts over water use or to large demonstrations, the social impact of drought is still poorly understood, and its economic cost even less so.
References
[7] AESN&DRIEE (2016), Mission sur le fonctionnement hydrologique du bassin de la Seine, Agence Eau Seine Normandie, https://doi.org/10.5194/hess-2016-308.
[18] Agence de l’Eau Seine Normandie (2016), Stratégie d’adaptation au changement climatique du Bassin Seine-Normandie, Comité de Bassin Seine-Normandie.
[66] Agence de l’Eau Seine-Normandie (2019), Etat des lieux du bassin versant Seine Normandie, https://www.eau-seine-normandie.fr/domaines-d-action/sdage/etat-des-lieux.
[33] Agreste (2022), Memento Ile-de-France 2022.
[13] Agreste (2021), L’occupation du sol entre 1982 et 2018, https://agreste.agriculture.gouv.fr/agreste-web/download/publication/publie/Dos2103/Dossiers%202021-3_TERUTI.pdf.
[28] Agreste (2011), Enquête sur les services publics de l’eau potable et de l’assainissement en 2008.
[54] APUR (2019), Les besoins en froid des bâtiments parisiens, https://www.apur.org/sites/default/files/documents/publication/etudes/note_149_besoins_froid_batiments_parisiens.pdf?token=s7MetKvA (accessed on 2 February 2022).
[16] Aquavesc, SEDIF, Sénéo, Ville de Paris (2020), Etude relative à la sécurisation de l’alimentation en eau potable de la zone interconnectée.
[38] Barcelona Metropolis (2023), Strategies to tackle water shortages, https://www.barcelona.cat/metropolis/en/contents/strategies-tackle-water-shortages#:~:text=However%2C%20Barcelona%20is%20one%20of,litres%20per%20person%20per%20day.
[85] Barton, J. and M. Rogerson (2017), “The importance of greenspace for mental health”, BJPsych. International, Vol. 14/4, pp. 79-81, https://doi.org/10.1192/s2056474000002051.
[50] BNPE (2022), Données de prélèvements France Entière, https://bnpe.eaufrance.fr/acces-donnees/france/annee/2021/exclureEcrasants.
[20] Boé, J. et al. (2018), Scénarios sécheresse sur le bassin Seine-Normandie, https://www.drieat.ile-de-france.developpement-durable.gouv.fr/IMG/pdf/rapport_boe_final.pdf.
[88] Bouscasse, H. et al. (2011), Evaluation économique des services rendus par les zones humides - Enseignements méthodologiques de monétarisation., irstea.
[86] Brander, L. et al. (2011), “Using Meta-Analysis and GIS for Value Transfer and Scaling Up: Valuing Climate Change Induced Losses of European Wetlands”, Environmental and Resource Economics, Vol. 52/3, pp. 395-413, https://doi.org/10.1007/s10640-011-9535-1.
[30] Chevrot, J. et al. (2018), “L’industrie francilienne: des mutations de long terme toujours à l’oeuvre”, INSEE Analyses 91, https://www.insee.fr/fr/statistiques/fichier/version-html/3673585/if_ina_91.pdf.
[76] Ciasca, B. et al. (2023), “Economic Cost of Drought and Potential Benefits of Investing in Nature-Based Solutions: A Case Study in São Paulo, Brazil”, Water, Vol. 15/3, p. 466, https://doi.org/10.3390/w15030466.
[46] Climate ADAPT (2018), Zaragoza: combining awareness raising and financial measures to enhance water efficiency, https://climate-adapt.eea.europa.eu/en/metadata/case-studies/zaragoza-combining-awareness-raising-and-financial-measures-to-enhance-water-efficiency/11210288.png/view.
[72] Conseil Régional d’Ile de France (2019), Stratégie régionale pour la biodiversité en Île-de-France, Région Ile de France.
[70] Conseil Régional d’Ile de France (2021), Plan régional pour une alimentation locale, durable et solidaire l’alimentation des franciliens : un enjeu de souveraineté, de santé et de relance, Région Ilede France.
[29] Cour des Comptes (2018), Alimentation en eau potable de la Métropole du Grand Paris.
[53] CPCU (2019), Informations et chiffres clés 2018, http://www.cpcu.fr/Nos-engagements/NOS-ENJEUX (accessed on 2 August 2022).
[42] Danva (2022), Water in figures, https://www.danva.dk/media/8746/5307102_water-in-figures-2022_web.pdf.
[2] DRIEAT (2023), Bilan annuel de l’étiage en Île-de-France - Saison 2022.
[51] DRIEAT Ile-de-France (2022), 30 millions de tonnes de déchets produits en Île-de-France, https://www.drieat.ile-de-france.developpement-durable.gouv.fr/30-millions-de-tonnes-de-dechets-produits-en-ile-a3857.html#:~:text=En%20%C3%8Ele%2Dde%2DFrance%2C,la%20chaleur%2C%20soit%20les%20deux. (accessed on 10 November 2023).
[84] DRIEAT Ile-de-France (2021), Approvisionnement et réseaux d’alimentation énergétique, https://www.drieat.ile-de-france.developpement-durable.gouv.fr/approvisionnement-et-reseaux-d-alimentation-a3427.html.
[41] Eau de Paris (n.d.), Calculer ma consommation d’eau, https://www.eaudeparis.fr/calculer-ma-consommation-deau.
[43] EPAL (n.d.), Efficient use, https://www.epal.pt/EPAL/en/menu/our-water/campaigns/efficient-use.
[9] EPTB Seine Grands Lacs (2022), Etude globale sur l’incidence socio-économique et environnementale des étiages sévères sur le bassin amont de la Seine, https://www.seinegrandslacs.fr/incidence-socio-economique-et-environnementale-des-etiages-severes.
[5] EPTB Seine Grands Lacs (2022), Etude sur l’incidence socio-économique et environnementale des étiages sévères.
[32] Fiorillo, D. et al. (2021), “Assessing the Impact of Climate Change on Future Water Demand using Weather Data”, Water Resources Management, Vol. 35, https://doi.org/10.1007/s11269-021-02789-4.
[37] France (2023), Données Vigi’Eau, https://vigieau.gouv.fr/.
[27] France (2023), https://lekiosque.finances.gouv.fr/site_fr/regions/data_region_produit_pays.asp?lreg=0300&typea=1.
[80] García-León, D., G. Standardi and A. Staccione (2021), “An integrated approach for the estimation of agricultural drought costs”, Land Use Policy, Vol. 100, https://doi.org/10.1016/j.landusepol.2020.104923.
[77] García-León, D., G. Standardi and A. Staccione (2021), “An integrated approach for the estimation of agricultural drought costs”, Land Use Policy, Vol. 100, p. 104923, https://doi.org/10.1016/j.landusepol.2020.104923.
[83] GIEC (2022), 6ème évaluation, Groupe de Travail 1: Les fondements scientifiques- Chapitre 8: Les changements des cycles de l’eau.
[81] GIEC (2022), 6ème Evaluation, Groupe de Travail 2: Impacts, Adaptation et Vulnérabilité, https://www.ipcc.ch/report/sixth-assessment-report-working-group-ii/ (accessed on 17 October 2022).
[60] Haropa Port (2023), Premier port fluvial européen, https://www.haropaport.com/fr/premier-port-fluvial-europeen (accessed on 10 November 2023).
[78] Horridge, M., J. Madden and G. Wittwer (2005), “The impact of the 2002–2003 drought on Australia”, Journal of Policy Modeling, Vol. 27/3, pp. 285-308, https://doi.org/10.1016/j.jpolmod.2005.01.008.
[79] Howitt, R. et al. (2015), “Economic Analysis of the 2015 Drought For California Agriculture”, University of California – Davis.
[56] INSEE (2023), Bilan economique 2022 - Ile-de-France, https://www.insee.fr/fr/statistiques/7619395?sommaire=7343425.
[23] INSEE (2023), Estimation de la population au 1ᵉʳ janvier 2023, https://www.insee.fr/fr/statistiques/1893198.
[34] INSEE (2023), https://www.insee.fr/fr/statistiques/1405599?geo=FRANCE-1.
[3] INSEE (2022), “Populations légales au 1er janvier 2020”, INSEE Flash Ile-de-France.
[25] INSEE (2022), “Produits intérieurs bruts régionaux de 2000 à 2020”, Insee Première, https://www.insee.fr/fr/statistiques/6440639#:~:text=9%20%25%20en%20volume-,Le%20PIB%20de%20l%27%C3%8Ele%E2%80%91de%E2%80%91France%20repr%C3%A9sente%20plus,%2DRh%C3%B4ne%2DAlpes%2012%20%25.
[65] INSEE (2022), Projections démographiques en Île-de-France à horizon 2070 : vieillissante, la région resterait la plus jeune de France métropolitaine, https://www.insee.fr/fr/statistiques/6666275# (accessed on 10 November 2023).
[31] INSEE (2022), “Projections démographiques en Ile-de-France à horizon 2070: vieillissante, la région resterait la plus jeune de France” INSEE Flash Ile-de-France n°72, https://www.insee.fr/fr/statistiques/6666275.
[24] INSEE (2021), “L’Île-de-France concentre la moitié des salariés du secteur privé les mieux rémunérés de France”, Insee Analyses Île-de-France, https://www.bnsp.insee.fr/ark:/12148/bc6p07b8dbq/f1.pdf.
[48] Institut d’aménagement et d’urbanisme (2016), Les lieux de l’industrie en Ile-de-France, https://www.institutparisregion.fr/fileadmin/NewEtudes/Etude_1244/Les_lieux_de_l_industrie_en_IDF.pdf (accessed on 8 September 2022).
[69] Institut Paris Region (2023), SDRIF-E 2040 - Un nouvel équilibre, https://www.institutparisregion.fr/planification/ile-de-france-2040/sdrif-e-2040/ (accessed on 10 November 2023).
[58] Institut Paris Region (2021), Fret et logistique dans l’Ouest francilien - État des lieux, Institut Paris Region.
[61] Institut Paris Région (2022), Vulnérabilités de l’Ile-de-France aux effets du changement climatique.
[4] Institut Paris Région (2021), Les Franciliens - Territoires et modes de vie.
[26] Institut Paris Région; INSEE; CCI (2021), Chiffres clés de la Région Ile-de-France.
[74] Jenkins, M., J. Lund and R. Howitt (2003), “Using economic loss functions to value urban water scarcity in California”, Journal / American Water Works Association, This article summarizes the methods and assumptions used to develop economic values of urban water use in California in 2020. <br/>Values in this study approximate how much the end users of urban water supplies are willing to pay for each additional unit of water along their demand curves<br/><br/>Good reference for ways of measuring water scarcity costs and cost of measures, https://doi.org/10.1002/j.1551-8833.2003.tb10292.x.
[11] Kapsambelis, D. (2018), Analyse des pertes de récoltes à l’échelle de l’exploitation agricole dans le cadre de l’assurance multirisques climatique en France métropolitaine.
[14] Kumari, U. et al. (2021), “Global water challenge and future perspective”, in Green Technologies for the Defluoridation of Water, Elsevier, https://doi.org/10.1016/b978-0-323-85768-0.00002-6.
[82] Lehner, F. et al. (2020), Partitioning climate projection uncertainty with multiple large ensembles and CMIP5/6, pp. 491-508.
[73] Logar, I. and J. van den Bergh (2012), “Methods to Assess Costs of Drought Damages and Policies for Drought Mitigation and Adaptation: Review and Recommendations”, Water Resources Management, Vol. 27/6, pp. 1707-1720, https://doi.org/10.1007/s11269-012-0119-9.
[39] Mayor of London (2022), Water use, https://www.london.gov.uk/who-we-are/what-london-assembly-does/questions-mayor/find-an-answer/water-use#:~:text=Figures%20from%20their%202021%2F22,of%20142.6%20litres%20per%20capita.
[10] Météo France (2022), Climat HD, Les tendances des évolutions du climat au XXème siècle, https://meteofrance.com/climathd.
[6] Météo France (2022), DRIAS Les futurs du climat.
[17] Météo France (n.d.), Climat HD, https://meteofrance.com/climathd (accessed on 10 November 2023).
[12] Météo Paris (n.d.), https://www.meteo-paris.com/actualites/la-secheresse-est-elle-aussi-inquietante-que-l-annee-derniere.
[19] Ministère de la Transition Écologique (2023), DRIAS - Les futurs du climat, http://www.drias-climat.fr/decouverte/cartezoom/experience/EUROCORDEX2020_DISTRIBUTION_ELAB/Q05/RCP4.5/RCP4.5/H2/APAV/APAV/A1 (accessed on 27 March 2023).
[62] Ministère de la Transition Ecologique et de la Cohésion des territoires (2021), Nouveau zonage d’exposition au retrait-gonflement des argiles : plus de 10,4 millions de maisons individuelles potentiellement très exposées, https://www.statistiques.developpement-durable.gouv.fr/nouveau-zonage-dexposition-au-retrait-gonflement-des-argiles-plus-de-104-millions-de-maisons.
[71] Ministère de l’écologie, du Développement et de l’Energie (2012), Synthèse du projet Explore 2070: Prospective socio-économique et démographique, https://professionnels.ofb.fr/sites/default/files/pdf/explore2070_4pages_prospective.pdf.
[57] Ministère des Finances (2023), “Le Chiffre du commerce extérieur”, Les chiffres clés de la région, https://lekiosque.finances.gouv.fr/fichiers/Etudes/Brochures/Reg_03.pdf.
[75] Naumann, G. et al. (2021), “Increased economic drought impacts in Europe with anthropogenic warming”, Nature Climate Change, quantitative assessment of future drought economic<br/>impacts across Europe.<br/><br/>EU economic, budgetary and demographic projections<br/><br/>Estimate future damages by building damage function <br/><br/>Damage function: D = S *(aP + b), where P = drought probability (return period) and S = country exposed sensitivity value (to drought)<br/><br/>Country sensitivity:, pp. 485-491, https://doi.org/10.1038/s41558-021-01044-3.
[47] OECD (2022), Israel’s sustainable water management plans, https://www.oecd.org/climate-action/ipac/practices/israel-s-sustainable-water-management-plans-d81db5f5/.
[8] PIREN-Seine (2018), Hydrologie du bassin de la Seine en quelques chiffres clefs, https://www.piren-seine.fr/sites/default/files/piren_documents/fiches_4_pages/Fiche%2520PS-Hydrologie_2109.pdf.
[63] Région Ile-de-France (2022), Plan de protection, de résistance et d’adaptation de la région île-de-france face au changement climatique.
[55] Région Ile-de-France (2021), Plan régional pour une alimentation locale, durable et solidaire.
[49] ROSE (2023), TABLEAU DE BORD, https://www.roseidf.org/panorama-regional-1-1-1/.
[52] RTE (2021), Bilan électrique en Île-de-France, https://assets.rte-france.com/prod/public/2022-07/Fiches%20Presse%20RTE%20-%20Bilan%20%C3%A9lectrique%202021%20IDF%20VDEF.pdf (accessed on 10 November 2023).
[36] SEDIF (n.d.), Passez à l’eau engagée, https://www.sedif.com/leauengagee#:~:text=Le%20saviez%2Dvous%20%3F,un%20foyer%20%C3%A0%20l%27autre.
[35] Sénéo; Aquavesc; SEDIF; Ville de Paris (2019), Etude relative à la sécurisation de l’approvisionnement en eau potable dans la zone interconnectée d’Ile-de-France.
[45] Toronto (n.d.), My Water Toronto, https://www.toronto.ca/services-payments/water-environment/how-to-use-less-water/mywatertoronto/#:~:text=In%202020%2C%20the%20average%20household,in%20your%20private%20plumbing%20system.
[15] United Nations office for disaster risk reduction (2021), Global Assessment Report on disaster risk reduction 2021: Special report on drought.
[1] Van der Schrier, G. et al. (2021), “The 1921 European drought: impacts, reconstruction and drivers”, Climate of the Past, Vol. 17/5, pp. 2201-2221, https://doi.org/10.5194/cp-17-2201-2021.
[44] Victoria State Government (n.d.), Target 150, https://www.water.vic.gov.au/liveable-cities-and-towns/using-water-wisely/target-150-target-your-water-use.
[22] Vidal, J. et al. (2010), “Multilevel and multiscale drought reanalysis over France with the Safran-Isba-Modcou hydrometeorological suite”, Hydrology and Earth System Sciences, Vol. 14/3, pp. 459-478, https://doi.org/10.5194/hess-14-459-2010.
[21] Ville de Paris (2021), Actualisation du diagnostic de vulnérabilités et de robustesse de Paris face aux changements climatiques et à la raréfaction de la ressource en eau.
[67] Ville de Paris (2019), Schéma directeur du réseau de froid parisien.
[64] Ville de Paris (2019), Stratégie de résilience.
[68] Voies Navigabels de France (2022), Les chiffres clés du transport fluvial – Bassin de la Seine, https://www.vnf.fr/vnf/chiffres-cles-fret-bassin-seine/#:~:text=Zoom%20sur%20le%20bassin%20de%20Seine&text=Avec%20plus%20de%2022%20millions,transport%C3%A9es)%20par%20rapport%20%C3%A0%202020. (accessed on 13 Décembre 2023).
[59] Voies Navigables de France (2022), https://www.vnf.fr/vnf/accueil/logistique-fluviale/economie-du-secteur-logistique/panorama-tendances-et-chiffres-cles-du-transport-fluvial/.
[40] Waternet (2021), Average water use household per day, https://www.waternet.nl/en/service-and-contact/tap-water/average-water-use/#:~:text=consumption%20per%20person.-,In%202021%2C%20the%20average%20water%20consumption%20of%20our%20customers%20was,liters%20per%20person%20per%20day.
[87] Woodward, R. and Y. Wui (2001), “The economic value of wetland services: a meta-analysis”, Ecological Economics, Vol. 37/2, pp. 257-270, https://doi.org/10.1016/s0921-8009(00)00276-7.
← 2. City of Paris website: https://www.paris.fr/pages/la-seine-2406
← 3. https://eau-grandsudouest.fr/usages-enjeux-eau/eau-grand-sud-ouest/etat-ressources-gestion-quantitative
← 4. Projections obtained from an average of the Special Report on Emissions Scenarios (SRES) reflecting different greenhouse gas emissions scenarios. The scenario chosen (RCP 4.5) represents an intermediate climate change scenario of a global average surface temperature increase of between 1°C and 4.5°C (compared with pre-industrial levels) by 2100. These projections cover the periods 2046-65 (2050 in the text) and 2070-2100 (2100 in the text). The hydrological trends deduced from these projections have been confirmed by more recent models (Agence de l'Eau Seine Normandie, 2016[18]).
← 5. SRES-A2 (slightly less pessimistic than RCP8.5).
← 6. For more information, see: https://www.drias-climat.fr/accompagnement/sections/310.
← 7. Fifty per cent of the region (western part, including Paris) is supplied with drinking water by the "interconnected" zone, in which the networks of the various water companies are interconnected.
← 8. In September 2003, the Neuilly-sur-Marne plant on the Marne River was subject to abstraction limits. In 2011 and 2012, the Arvigny plant on the Champigny aquifer in Seine-et-Marne experienced abstraction limits (5% of the drinking water company's production was affected).
← 9. These consumption data cover the area served by SEDIF and Eau de Paris, in the absence of data on per capita consumption for the whole of the region.
← 11. Eau Île-de-France: https://eau-iledefrance.fr/grigny-coca-cola-a-un-coup-de-pompe/
← 12. Figures provided by the regional Region Chamber of Agriculture: https://idf.chambre-agriculture.fr/la-chambre/lagriculture-francilienne/chiffres-cles/#:~:text=1er%20bassin%20national%20de%20consommation,22%25%20%C3%A0%20l%27export.
← 13. Between 2010 and 2020, irrigable land increased by 7%, with a 22% increase expected by 2050. As the study aims to assess the impact of restricting irrigation during periods of severe soil dryness and hydrological drought, irrigated areas have been used to analyse crop irrigation requirements.
← 14. The 23 European countries are: Austria, Belgium, Denmark, Estonia, Finland, France, Germany, Greece, Hungary, Ireland, Italy, Latvia, Lithuania, Luxembourg, Malta, Netherlands, Poland, Portugal, Slovakia, Slovenia, Spain, Sweden and United Kingdom.
← 16. Decree dated 24/08/17 modifying, through a series of ministerial decrees, the provisions relating to the discharge of hazardous substances into water from installations classified for the protection of the environment.
← 17. The ESVD database can be accessed at: https://www.esvd.net/.