All official European Union website addresses are in the europa.eu domain.
See all EU institutions and bodiesDo something for our planet, print this page only if needed. Even a small action can make an enormous difference when millions of people do it!
Indicator Specification
This item is open for comments. Login with your Eionet account in order to see and add comments. See comments section below
Drought is a recurring and extreme climate event that is induced by a temporary water deficit and may be related to a lack of precipitation, soil moisture, streamflow, or any combination of the three taking place at the same time (Wilhite and Glantz 1985). Drought differs from other extreme natural events in several ways. First, unlike earthquakes, floods or tsunamis that occur along generally well-defined fault lines, river valleys or coastlines, drought can occur anywhere (with the exception of desert regions where it does not have meaning) (Goddard et al. 2003). Secondly, drought develops slowly, resulting from a prolonged period (from months to years) of water supply conditions that are below the average at a specific location (Dracup et al. 1980).
Because of their long-lasting socioeconomic impacts, droughts are by far considered the most damaging of natural disasters (Sivakumar et al. 2014). The immediate impacts of short-term droughts (i.e. a few weeks duration) are, for example, a fall in crop production, poor pasture growth and a decline in fodder supplies from crop residues. Prolonged water shortages (e.g. of several months or years duration) may, among others, lead to a reduction on hydro-electrical production and potentially increase wildfire occurrences on land-based natural and managed ecosystems (Mishra and Singh 2009).
Although droughts are typically associated with aridity (Seager et al., 2007, Güneralp et al., 2015), they can occur in most parts of the world, even in wet and humid regions (Ivits et al., 2013; Lewinska et al., 2018), and can profoundly impact agriculture, basic household welfare, tourism, ecosystems and the services they provide (Goddard et al., 2003, Dai, 2011). For example, in arid and semi-arid ecosystems (including the Mediterranean regions), limiting water availability is a recurrent phenomenon and governs plant growth and phenology (Reichstein et al., 2002). On the other hand, in temperate, boreal and tropical ecosystems, sporadic prolonged dry periods can lead to water-limited conditions and can have far-reaching impacts on ecosystem carbon (C) balance (Ciais et al., 2005; Granier et al., 2007; Doughty et al., 2015) and structure (Orth et al., 2016).
The monitoring and assessment of drought impacts is complex because different types of impacts vary in their intensity, often in different phases of the given drought event, as indicated above. Therefore, most empirical studies of drought impacts have focused on agricultural crop production, which is direct, immediately observable, well understood, and easy to quantify (Wilhite, 2000; Ding et al., 2011). Reports about drought impacts in the category 'terrestrial ecosystems' were only found for a few years and are limited in number in the EDII database (Stahl et al., 2016). This agrees with the earlier conclusions of Lackstrom et al. (2013), which claim that there is a lack of data and understanding of drought impacts on sectors other than agriculture and water resources.
Differences in the physiological response of vegetation to water deficits determine different levels of sensitivity and resilience of terrestrial ecosystems to drought (Chaves et al. 2003, McDowell et al. 2008), and ultimately influence the type of drought impacts, i.e. differentiating those impacts that slow growth (Pasho et al. 2011) or reduce greenness (Ji and Peters 2003), those that lead to loss of biomass (Ciais et al. 2005), and those that result in plant mortality (Allen et al. 2010, Adams et al. 2009). Consequently, significant changes in the amount of vegetation productivity provide an indication/early warning of imminent irreversible impacts in ecosystems' equilibrium states (De Keersmaecker et al., 2015).
The indicator addresses anomalies and long term trends of vegetation productivity derived from remote sensing observed time series of vegetation indices in areas that are pressured by drought.
Drought pressure is computed as the combination of significant Pearson correlation coefficients (r) derived between time series anomalies of yearly vegetation productivity and the anomalies of various drought hazards (dH) during the growing season.
Drought impact is indicated as the most sever negative productivity anomaly under drought pressured areas as well as long term decreasing linear trends of annual vegetation productivity for areas that are pressured by drought hazards. Detailed indicators’ specifications are presented under Methodology.
Area of drought impact (km2);
Normalized vegetation productivity anomalies (standard deviation);
Long term linear trend(%)
Environmental policy is highly dependent on the monitoring and assessment of its targets, be it land take, biodiversity loss, environmental pollution or land degradation. In May 2011, the EU adopted a Biodiversity Strategy to 2020, which identifies 6 priority targets and 20 actions for the European Union to reach the target of "Halting the loss of biodiversity and the degradation of ecosystem services in the EU by 2020, and restoring them in so far as feasible, while stepping up the EU contribution to averting global biodiversity loss". For the EU, the opportunity cost of not reaching the 2020 biodiversity headline target of halting the loss of biodiversity and ecosystem services has been estimated at 50 billion EUR per year[1]. In addition, to undermining these economic benefits, loss of biodiversity means that ecosystems and societies that rely upon them are more fragile and less resilient in the face of challenges such as climate change, pollution and habitat destruction. Droughts have an impact on several land and soil functions, as well as ecosystem services, both in urban and rural areas. For example, droughts have an impact on water resources available for human use in agriculture, cause habitat loss, migration of local species and their replacement by alien ones in open rural systems, and consequently soil erosion and biodiversity degradation. By pressuring natural ecosystems, droughts hamper the achievement of EU Biodiversity 2020 objectives.
Drought pressure on natural ecosystems has also an important role on the implementation of the EU Strategy on Green Infrastructure (GI). In contrast to the most common ‘grey’ (man-made, constructed) infrastructure approaches that serve one single objective, GI promotes multifunctionality, which means that the same area of land is able to perform several functions and offer multiple benefits if its ecosystems are in a healthy state. More specifically, GI aims to enhance nature's ability to deliver multiple valuable ecosystem goods and services, potentially providing a wide range of environmental, social, climate change adaptation and mitigation, and biodiversity benefits. Drought diminishes the normal condition of ecosystems and their capacity to provide services that could be integrated in green infrastructures.
Under EU legislation adopted in May 2018, EU Member States have to ensure that greenhouse gas emissions from land use, land use change and forestry (LULUCF) are offset by at least an equivalent removal of CO₂ from the atmosphere in the period 2021 to 2030. Ultimately, the capacity of forests and soils on a given area of land to remove carbon from the atmosphere will depend on a number of natural (regional/geographical) circumstances such as variations in growing conditions (temperature, precipitation and droughts) and natural disturbances (storms, fires) as well as past and present management practices (e.g. rotation lengths which affect the distribution of age classes in forest stands). By measuring changes in emissions and removals relative to business-as-usual projections, these circumstances (such as drought pressure) will be "factored out" so that only changes related directly human-induced activities are measured. This also provides incentives for improving on the current situation and gives an equal value to mitigation whether through sequestration or conservation or material and energy substitution.
The role of the CAP is to provide a policy framework that supports and encourages producers to address economic, environmental (i.e. relating to resource efficiency, soil and water quality and threats to habitats and biodiversity) and territorial challenges, while remaining coherent with other EU policies. This translates into three long-term CAP objectives: viable food production, sustainable management of natural resources and climate action and balanced territorial development. Given the pressure of drought on natural resources, agriculture has to improve its environmental performance through more sustainable production methods. Farmers have also to adapt to challenges stemming from changes to the climate by pursuing climate change mitigation and adaption actions (e.g. by developing greater resilience to disasters such as flooding, drought and fire). Understanding the spatio-temporal distribution of drought pressures on land, will contribute to a better, faster and more informed implementation of CAP reforms and improve the quality of life of rural populations in Europe.
No specific targets.
Drought pressure is derived at the pixel level and is based on the combination of Pearson correlation values (r) computed for statistically significant (p-value < 0.05) regression coefficients of three Ordinary Least Squares (OLS) models. The estimated OLS models fit yearly anomalies of vegetation productivity to the severity of three drought hazard indicators (i.e. Soil Moisture, SPI-3 and -12 - see under Data specifications) occurring during the growing seasons from the years 2000 to 2016. Individual OLS models, commonly known as linear regressions, are computed for each individual grid-cell in the raster database, as follows:
Y=B01+B1S1+e1
Y=B02+B2S2+e2
Y=B03+B3S3+e3
where Y denotes the yearly anomalies of vegetation productivity, B01-B03 are the intercept terms, B1-B3 are the regression coefficients, S1-S3 are the drought severity indices during the growing season (GS), and e1-e3 are the error terms of the models. As defined by Spinoni et al. (2014), drought severity can be computed as the sum of negative drought Hazard (dH) values (represented by the bars in red to yellow colours in the plot below) from the start to the end month of the vegetation GS, as follows:
S=∑gs Xm
where Xm is a negative drought Hazard, dH, intensity value occurring in month m within the GS. Severity should not be mistaken for intensity, which is usually referred to the dH value in each month during the GS.
Figure 1. Timeseries of monthly dH (i.e. SPI-3) and changes in the start and length of annual GS periods for a single grid-cell in the raster database (adapted from a report of the JRC Global Drought Observatory: https://edo.jrc.ec.europa.eu/gdo/).
Input data for estimating individual OLS models, i.e. the yearly vegetation productivity and the S values during GS, were first de-trended so as to avoid exaggerated statistically significant relations from spurious regressions (Granger and Newbold 1974). Detrending was performed by substituting the original values of both independent (S) and dependent (Y) variables by the residuals of their respective linear trend estimations. The significance of the OLS models was based on the p-value of the Student’s T-Test statistic; whenever the p-value < 0.05, then the of the respective model was assumed to be statistically significant.
The correlation coefficients of the drought pressure index are selected as the maximum r value among all significant and positive regression coefficients (B1-B2) of the individual OLS models. Since an increase of drought hazard severity is expected to increase vegetation stress conditions, then only positive correlation coefficients were taken into account in the final indicator. Negative correlation coefficients might be indicative of management activities, such as irrigation practices, which are not addressed by the indicator. Moreover, positive correlation coefficients were kept only when all regression coefficients for the three OLS models were positive.
The proposed approach follows that of Vicente-Serrano et al. (2013), who mapped the response of vegetation to drought across global land biomes according to the highest correlation coefficient found between monthly GIMMS-NDVI and SPEI values measured at time-scales varying from 1- to 24-months. It was shown that different ecosystems react differently to the type and temporal scale of drought hazard values during the GS. Some vegetation types respond immediately to a short lack of precipitation (as measured by SPI-3), while other ecosystems respond more intensively to anomalies in the soil moisture content (anomSM). Moreover, Hofer et al. (2012) have shown that perennial vegetation types in the Iberian Peninsula, such as forests, are less sensitive to short term water deficits, thus requiring an evaluation of the association with large scale drought indices, such SPI-12. Similar conclusions were also drawn for the Iberian Peninsula by Gouveia et al. (2016) and at the global level by Vicente-Serrano et al. (2015). These studies have shown that by consequence of different physiological, anatomical or edaphic factors, some vegetation communities show a response to short periods of water deficit, whereas others may be resistant to soil water deficit, and respond to longer drought time scales representative of water deficits of longer duration.
Drought impact was quantified as:
-the per pixel minimum of the standardized vegetation productivity anomaly during 2000-2016;
-timing of the minimum anomaly, which records the year of the most sever drought impact on vegetation productivity;
-long term significant negative trends in the yearly vegetation productivity indicating lasting impact of intense drought events.
When analyzing drought impact, only those pixels were taken into account which were under drought pressure, i.e. where the regression among the drought indices and the productivity anomalies were significant and positive. Moreover, to account for outliers in the trends, which are occurring due to erroneous satellite image measurements or human land use activities in drought affected areas (such as irrigation or clear cuts), it was decided to remove all trend values below the 10th and above the 90th percentiles. Finally, to avoid the inclusion of trends derived from spurious regressions, only statistically significant (p-value < 0.05) negative trends were kept as part of the final indicator. The significance of the trends was based on the Getis-Ord Gi inferential statistic (Getis and Ord, 1992).
All input dataset are derived from global sources with a wall-to-wall coverage of the land surface. No gap filling was needed.
The approach cannot account for land use/land cover changes that have occurred within a pixel for the period of analysis. For example, clear cuts within forest ecosystems or the use of irrigation systems as part of the management in agricultural areas might increase or decrease the vegetation productivity independently of drought occurrences. This can introduce noise in the datasets that might bias the pixel-based relationships between drought pressure and vegetation productivity.
Another source of uncertainty is related with the simplification of the drought impact model for its implementation in the operational setting. In this study, only meteorological drought distribution and intensity is considered. Still, in some cases, the start, end, severity, and spatial extent of a drought, as well as the propagation of its impacts through the whole land systems might be changing due to additional climate and/or surrounding biophysical conditions, such temperature, snowpack, albedo and soil water holding capacity.
The dataset represents the average trend of productivity of all terrestrial ecosystems within an area covered by a pixel of 500x500m. Therefore, the dataset can only be used at the ecosystem level indicating drought impacts on main terrestrial ecosystems. As opposed to field measurements, remote sensing products measure vegetation light absorption from a satellite at several hundred km height which might introduce bias due to atmospheric disturbances.
No uncertainty has beenidentified
Work specified here requires to be completed within 1 year from now.
Work specified here will require more than 1 year (from now) to be completed.
For references, please go to https://www.eea.europa.eu/data-and-maps/indicators/drought-impact-on-vegetation-productivity or scan the QR code.
PDF generated on 23 Sep 2024, 08:12 PM
Engineered by: EEA Web Team
Software updated on 26 September 2023 08:13 from version 23.8.18
Software version: EEA Plone KGS 23.9.14
Document Actions
Share with others