1. Introduction
The water stored on land surfaces, referred to as terrestrial water storage (TWS), plays a key role in the hydrological cycle. In particular, soil moisture is an essential element contributing to land–atmosphere coupling (e.g., Koster et al. 2004) and is known to be important for numerical weather prediction (e.g., Beljaars et al. 1996) and seasonal forecasting (e.g., Koster et al. 2000), as well as climate modeling in general (e.g., Shukla and Mintz 1982; Milly and Dunne 1994). The soil moisture–precipitation feedback (e.g., Betts et al. 1996; Eltahir 1998; Schär et al. 1999) also demonstrates its relevance for local and regional climate. An accurate understanding of such processes, as well as the ability to simulate present-day soil moisture evolution on continental to subcontinental scales, is an important prerequisite for predicting future changes in soil moisture and related impacts on the occurrence of floods (e.g., Milly et al. 2002; Kleinn et al. 2005) or droughts (e.g., Wetherald and Manabe 1999; Seneviratne et al. 2002; Schär et al. 2004b).
Reliable datasets of TWS and its components (mostly soil moisture, snow, groundwater, and surface water) are extremely scarce, however. Soil moisture is measured routinely only on a few spots around the world, mostly in the former Soviet Union, Mongolia, China, India, and Illinois (Robock et al. 2000). Model-derived estimates (e.g., Mintz and Serafini 1992; Dirmeyer et al. 1999, 2002) are available, but the uncertainty associated with these products is relatively large (e.g., Entin et al. 1999). Snow and groundwater measurements tend for their part to be even more limited (e.g., Rodell and Famiglietti 2001; Seneviratne et al. 2004). Remote sensing holds quite some promise in this area, in particular the recently launched Gravity Recovery and Climate Experiment, which appears capable of successfully capturing large-scale variations in TWS (Tapley et al. 2004; Wahr et al. 2004). However, this mission has only recently started and cannot provide any information on past values of TWS. In general, large-scale soil moisture estimates from different sources substantially disagree (Reichle et al. 2004).
In a recent paper, Seneviratne et al. (2004, hereafter referred to as S04) have demonstrated that retrospective basin-scale estimates of TWS can be successfully derived from combined atmospheric and terrestrial water-balance computations, using streamflow measurements and reanalysis data. Their study focused on the Mississippi River basin for the time period 1987–96 and used European Centre for Medium-Range Weather Forecasts (ECMWF) 40-yr Re-Analysis (ERA-40) data in combination with U.S. Geological Survey (USGS) streamflow measurements. In particular, estimates derived with this approach for Illinois (∼2 × 105 km2) were found to agree very well with in situ observations of soil moisture, groundwater, and snow.
The present study extends this previous analysis to 37 river basins in Europe, Asia, North America, and Australia and has as its main aim the creation of a basin-scale dataset of monthly TWS variations covering up to 40 yr. The described dataset is freely available on the Internet (see section 6) and is expected to be of particular use for climate applications in regions where large-scale observational data are not available.
The paper is structured as follows. Section 2 gives an overview of the methodology and describes the river basins investigated as well as the datasets employed. Section 3 presents the main results of this study, that is, the climatologies of the derived estimates for the considered river basins. In section 4, the ERA-40 moisture flux convergence and the derived water-balance estimates are compared against historical data, soil moisture and snow depth observations, and the simulated ERA-40 TWS. Section 5 presents a possible application of the computed dataset for climate model validation, followed by information for data download (section 6). Finally, the last section provides a summary of the results as well as the main conclusions of this study.
2. Computation of the water-balance estimates
a. The water-balance equations
b. Investigated river basins
The combined water-balance Eq. (3) is used here for the investigation of 37 river basins in Europe, Asia, North America, and Australia. The analysis covers the 44-yr time period of the ERA-40 reanalysis but is temporally limited in some river basins depending on the availability of runoff data. An overview of the geographical location of the computed catchment areas can be found in Fig. 1. The size of the river basins range from 3 × 104 up to 3 × 106 km2 (see Table 1). Note that the critical size for water-balance computations using high-resolution reanalysis data is supposed to be in the order of 105 km2 (Yeh et al. 1998; Berbery and Rasmusson 1999; S04). To reach this threshold, some smaller basins in Europe have been combined into larger compound domains for some of the analysis (i.e., the French basins and central European basins).
c. Employed datasets
1) Streamflow data
The monthly streamflow data has been obtained from the Global Runoff Data Centre (GRDC), from the USGS, or from local sources in the respective countries (see Table 1). In each river basin (see Fig. 1), the most downstream runoff station has been considered for the computation of the water balances in order to maximize the area of the investigated region. Additionally, some subbasins of large Asian and North American river basins were also included in the analysis. For many rivers (especially in Russia), it was difficult or even impossible to get more recent runoff data (after 1990).
2) ERA-40 moisture flux divergence and changes in atmospheric moisture content
The vertically integrated moisture fluxes and changes in atmospheric moisture content are taken from the ERA-40 reanalysis product (Simmons and Gibson 2000; Simmons et al. 2004) of the ECMWF. The ERA-40 reanalysis covers the period from mid-1957 to mid-2002; here we use data for the years 1958–2001, that is, 44 yr altogether.
For a detailed description of the fields used from ERA-40 as well as the computation of the moisture flux divergence, the reader is referred to S04. The ERA-40 model has a T159 spherical harmonic representation for the atmospheric dynamic and thermodynamic fields, and uses a reduced Gaussian grid (Hortal and Simmons 1991) with grid spacing of 112 km. There are 60 levels in the vertical, with a particular high resolution in the lower troposphere: The lowest model level is at 10 m above the surface and there are 8, 11, 15, 17 and 22 levels below 500, 1000, 2000, 3000, and 5000 m, respectively. The reanalysis uses a three-dimensional variational assimilation system (Courtier et al. 1998) with a 6-h analysis cycle. Additional documentation on ERA-40 can be found at http://www.ecmwf.int/research/era/.
The two ERA-40 fields used in the water-balance computations (i.e., moisture flux divergence and changes in atmospheric moisture content) contain assimilated humidity and wind observations. These fields are averaged over the basins using the fractional coverage of the catchments in each ERA-40 grid box as a weighting factor [following Schär et al. (2004a), adapted for the reduced Gaussian grid].
3) Topographical data
Seneviratne et al. (2004) used simple quadrilaterals to define the area of the river basins for the computation of the terms in (3). Here, the catchments' definitions are derived from the HYDRO1k topographical dataset (for data download, see http://edc.usgs.gov/products/elevation/gtopo30/hydro/index.html). HYDRO1k, developed at the USGS Earth Resources Observation Systems (EROS) Data Center, is a geographic database providing a comprehensive and consistent global coverage. The basis of all data layers available in the HYDRO1k database is the hydrologically corrected digital elevation model (DEM), which is based on the GTOPO30 dataset, a 30 arc s DEM of the world. To ensure that the DEM is able to reproduce the correct movement of water across its surface, the DEM was processed to remove elevation anomalies that can interfere with hydrologically correct flow.
For the Murray–Darling basin in Australia, the global Simulated Topological Network at 30-min spatial resolution (STN-30; Vörösmarty et al. 2000) is used to define the catchment, since there is no coverage of the HYDRO1k dataset on this continent. STN-30 represents rivers as a set of spatial and tabular data layers derived from a 30-min flow-direction grid (for data download, see http://www.watsys.sr.unh.edu/Stn-30/stn-30.html).
3. Results
a. Introductory examples and drift correction
This section exemplifies the analysis of the diagnosed monthly variations in TWS for the Ob and the Rhine River basins.
Figures 2a and 2b display the mean annual cycle of the moisture flux convergence, surface runoff, and the diagnosed monthly variations in TWS (raw and drift-corrected values; see below) for the Ob and the Rhine River basins. Not shown in these panels are the monthly changes in atmospheric moisture content {
The Ob basin displays a runoff peak in June, which is produced by snowmelt (Fig. 2a). The Rhine runoff (Fig. 2b) does not exhibit a pronounced seasonal cycle, due to compensating characteristics in the upper and lower reaches during winter and summer (winter with high discharge in low regions and storage of water in mountains as snow; spring/summer with low discharge in lower regions and increased runoff in elevated areas due to snowmelt). For both basins, there is atmospheric moisture convergence in the cold season and moisture divergence in summer.
As discussed in S04, the drift in TWS (uncorrected full lines in Figs. 2c,d), which is not homogeneously distributed over the entire time period and can even change sign within the time period considered, is likely due to artificial drifts in the atmospheric moisture convergence data. Indeed, errors in runoff measurements are expected to be small. Accordingly, inhomogeneities are seen in the cumulated water vapor flux convergence (see Figs. 2c,d, long-dashed lines), whereas the observed river discharge appears homogeneous (see Figs. 2c,d, short-dashed lines). Possible explanations for the existence of drifts in the atmospheric moisture convergence data are discussed in more detail in section 4b). Overall, it is unlikely that these drifts in the water balances correspond to actual variations in TWS, though the latter can be important in some regions and could contribute to part of the signal (see section 4b). Because of the lack of information on such natural sources of drifts, the most appropriate procedure is to assume that the observed drifts are purely artificial.
As the drift of {
The impact of the drift correction is shown in Figs. 2a and 2b by the thin (uncorrected) and bold (corrected) solid lines. In the case of the Ob basin, the corrections are small and the two lines coincide. In the case of the Rhine, the corrections are substantial and amount to ∼0.15 mm day−1.
b. Regional climatologies
Figures 3 –5 display the mean annual cycle of the water-balance quantities for most considered European, Asian, North American, and Australian river basins in the same format as Figs. 2a and 2b, averaged over the denoted years. Some smaller basins in Europe have been combined into one larger compound basin. For instance, the French basins refer to the Seine, Loire, Garonne, and Rhone (see Fig. 1).
Beside the monthly variations in TWS, we have also computed “absolute” values of TWS for all basins using (4) with drift-corrected {
1) European river basins
Most of the European rivers (central European and French basins, Danube, Wisla, Odra, Rhone, Seine) do not exhibit a pronounced seasonal cycle of runoff. Normally, runoff is lowest in summer and highest in (late) winter. The Po basin shows two runoff maxima associated with the two precipitation maxima in spring and late autumn (Frei and Schär 1998). The two precipitation maxima are also evident in the atmospheric moisture convergence. In most other basins of Europe, there is atmospheric moisture convergence in the cold season and moisture divergence in summer.
2) Asian river basins
The Lena, Yenisei (not displayed), and Volga River basins in northern Asia show a seasonal cycle that is highly dependent on the runoff peak in May or June, which is generated by snowmelt (e.g., Dettinger and Diaz 2000). The Volga basin displays moisture divergence in summer, which contributes to the TWS depletion. The Lena and Yenisei River basins display moisture convergence during the whole year with maximum values in spring and fall, and the TWS depletion in these basins is mostly determined by snowmelt and runoff.
The two central Asian river basins Amudarya and Syrdarya (not shown) display a similar but less pronounced snowmelt peak in runoff in June and July, but the TWS variations are more affected by moisture convergence in winter and divergence in summer, as is typical for a semiarid region.
The Amur and the Changjiang basins exhibit a strong moisture convergence in summer associated with the monsoon occurring in these regions. This is also influencing river runoff, which shows an increasing tendency until the end of the summer.
The Don, Dnepr, and Neva basins show small monthly variations in runoff. Streamflow values of the Don are very low and the variations in TWS follow closely the moisture convergence. In the case of the Neva, runoff is relatively large, but without a snowmelt peak similar to the other higher-latitude river basins (Volga, Ob, Lena, and Yenisei). This is partly explained by lakes that are damping the effect of spring snowmelt (Pardé 1947).
3) North American and Australian river basins
The seasonal cycles of the Columbia, Yukon, and Mackenzie River basins in North America show a runoff peak in June, which is generated by snowmelt. The Yukon and Mackenzie basins exhibit moisture convergence during the whole year and the water storage depletion in this region is mostly driven by changes in snow cover (same pattern as in the Lena and Yenisei River basins). The variations in TWS in the Columbia region is strongly dependent on the seasonal development of the moisture convergence, which displays high values in autumn/winter, and large divergence in summer. The western Mississippi subbasins (Missouri and Arkansas) show small and constant runoff values throughout the year, whereas the eastern subbasin (Ohio) displays a runoff peak in March. All subbasins of the Mississippi exhibit moisture divergence in summer.
The Murray–Darling basin in Australia displays only very small runoff values throughout the year (note the differing scale of the y axis). Thus the variation in TWS is defined by the atmospheric moisture flux, which shows convergence in austral winter and divergence in summer.
4. Discussion and comparison with other datasets
The combined water-balance method used here for the diagnosis of monthly variations in TWS has been validated by S04 for the fraction of Mississippi basin covering Illinois for the time frame 1987–96. The analysis showed excellent agreement with observational data of soil moisture, groundwater, and snow from the Illinois State Water Survey (ISWS) and the Midwest Regional Climate Center (MRCC).
Here we provide some additional validation and comparisons against other datasets. For all the analysis, the employed runoff data can be assumed to be accurate within a few percent for long-term averages. Errors in runoff observations should thus not contaminate the results of the water-balance computations. Observational runoff errors range from less than 5% to 5%–10% and depend on the instrumentation and the methods used to distribute discharge data both in time and space (Winter 1981; Rantz 1982a, b). When analyzing model-derived TWS variations against our estimates, one should take into account that the measured river discharge may be anthropogenically influenced as a result of water management procedures (Hagemann and Dümenil 1998).
a. Comparison with historical values of P − E in Europe
The ERA-40 water vapor flux convergence is compared here with water-balance estimates from several historical studies. Alestalo (1983) calculated vertically integrated, annual, and monthly averaged divergences of the horizontal water vapor flux directly from aerological observations for two European domains. The corners of his domains were defined by World Meteorological Organization (WMO) aerological stations (see Fig. 6). Korzun (1974) as well as Baumgartner and Reichel (1975) mapped the world water balances and estimated precipitation (P), evaporation (E), and runoff by using long-term control relationships (e.g., P – E = runoff). In his investigations on the heat balance of the earth, Budyko (1963) drew conclusions on the evapotranspiration in various regions. Combining these estimates with precipitation and runoff data, he compiled a picture of the global water cycle. Alestalo (1983) used the aforementioned three studies to estimate the water-balance components for the above-mentioned European domains (see Table 2).
According to (2), − {
For the smaller central European region [region M in Alestalo (1983); see Fig. 6], however, ERA-40 underestimates the moisture convergence by up to 50% compared to the other estimates. This is a striking result, since the same observations used in Alestalo (1983) are also assimilated in ERA-40. It is likely that this underestimation is very dependent on the definition of the region M, and in particular its southern border that runs parallel to the Alpine ridgeline. When extending the area slightly to the south, the ERA-40 convergence increases significantly and agrees much better with the historical estimates. This indicates the difficulty when averaging the ERA-40 fields over small areas in the vicinity of complex topography. Although each grid box is included in the computation of the estimates according to its fractional coverage of the domain [see section 2c(2)], sub-grid-scale meteorological features, which are especially strong in a topographically complex region such as central Europe, are not fully represented, leading to unclosed water budgets. The use of physical catchments as the basis for the domain definitions, as done in our study, is here of some advantage.
Nevertheless, Table 2 suggests that ERA-40 has a tendency to underestimate atmospheric water vapor convergence, and this can at least partly explain the diagnosed negative long-term drifts in water storage (see Fig. 2d and next subsection) in many smaller river basins of this region (Danube, Wisla, Odra, Elbe, Weser, Rhine, Rhone, Loire, Garonne, Ebro).
b. Long-term balance of the moisture flux convergence and runoff
To validate the long-term moisture flux convergence with runoff data (note that the contribution of the change in atmospheric water vapor to the drift in TWS is negligible, since its long-term mean is almost zero; see previous section), Fig. 7 shows the averages of the drift in water storage over the computed years {
However, the water storage in the basins of western Russia and northern and central Asia is quite well balanced not only for the large regions, but also for most small basins (Don, Dnepr, Syrdarya, subbasins of the Ob, Yenisei, and Lena), whereas many, even large central and western European basins, exhibit a pronounced (mostly negative) drift in long-term water storage (Danube, Wisla, Odra, Elbe, Weser, Rhine, Rhone, Loire, Garonne, Ebro). In North America, the water balance is stable for the large river basins (Mississippi and Mackenzie), whereas most small domains are not well balanced (with the exception of the Yukon basin). This analysis suggests that the magnitude of the imbalances is not only dependent on the size of the river basins but also on the geographical region under consideration (i.e., climatological and topographical properties, e.g., Seneviratne 2003).
As previously discussed in S04 (see also Figs. 2c,d), the drift in the integrated TWS is often not homogeneously distributed over time and can even change sign within the analysis period. This is likely due to changes in the quality and quantity of observations entering the ERA-40 data assimilation system (Betts et al. 2003b), for instance, the introduction of satellite data in the late 1970s, the diminution of the radiosonde network in the 1990s in the countries of the former Soviet Union, Africa, and South America, and the increasing number of satellite observations in the same period. More detailed studies on the influence of changes in the global observing system on the ERA-40 hydrological cycle and on trends in this dataset have been published recently. Bengtsson et al. (2004a) conclude that changes in the observing systems, which have taken place in the last 40 yr, have introduced artificial climate trends of similar magnitude as the real ones in the reanalysis data. In particular, they find an overestimation of the trend in integrated atmospheric water vapor, which is likely to affect the whole water cycle. Bengtsson et al. (2004b) further demonstrate that the hydrological cycle of ERA-40 is largely determined by observations of dynamical quantities and the assimilation system's model physics, while humidity observations appear to have a comparatively small impact.
An analysis in the Rhine basin shows that when the Alpine region is omitted from the analysis to reduce the impact of glaciers and complex topography (by considering the basin downstream of Basel), the negative drift in water storage is significantly reduced (from −0.195 to −0.032 mm day−1 for the period 1961–93; see Fig. 2d), while the overall pattern of the drift persists. This confirms the findings made in the previous section and shows again the difficulty when averaging the ERA-40 convergence over the topographically complex Alpine region. The grid resolution of ERA-40 is too small to resolve the steep orography of the Alps and the involved convective precipitation events. However, this cannot serve as an explanation for the imbalances in other basins without complex topography (e.g., Elbe). The reason for the drifts in these basins must be associated with quality issues of the ERA-40 water budget unaffected by limited horizontal resolution.
S04 found similar imbalances in water storage of the different Mississippi subbasins using the same computational method and reanalysis dataset. Gutowski et al. (1997) concluded in their comparisons of National Centers for Environmental Prediction (NCEP)–National Center for Atmospheric Research (NCAR) reanalysis water vapor convergence with river discharge in the same region that the obtained drift is largely due to inaccuracies in atmospheric transport. This seems to be a common feature of the combined water-balance method and has also been seen in calculations with raw radiosonde data (Rasmusson 1968) and other reanalysis data (Oki et al. 1995).
Beside artificial drifts in the convergence from the reanalysis, possible “natural” reasons for the drift in TWS can be anthropogenic groundwater use. According to Zektser et al. (2005), the large demand for usable water in combination with the semiarid climate has led to groundwater overdraft in the southwestern United States. This produced declining trends in aquifer storage and hydraulic height in this region. This decline can be seen also in some groundwater level records from the USGS (e.g., see, http://nwis.waterdata.usgs.gov/nwis/gwlevels/?site_no=393403101575901; a station in the Arkansas River basin with an average groundwater level decrease of 0.5 mm day−1 over the last 30 yr). Since this anthropogenic removal of water is not accounted for in the water-balance calculations, it possibly can explain the occurring positive drift in computed TWS in river basins of this region (i.e., Missouri, Arkansas). Other neglected factors possibly contributing to a drift in the calculation are direct groundwater discharge to the ocean (which is not included in runoff and represents approximatively 6% of the total annual global water gain by the oceans with highly variable spatial distribution; Zektser and Loaiciga 1993) and interbasin groundwater flow.
As described in the previous section, we correct the estimates of TWS variations by high-pass filtering (see Figs. 3 –5). This approach provides more realistic estimates than the raw diagnostic values, but some open questions remain. In particular, high-pass filtering would be of limited value if the drift of the assimilation system should have a strong seasonal cycle.
c. Correlation with soil moisture and snow depth observations
Soil moisture measurements from the Global Soil Moisture Data Bank (Robock et al. 2000) available for the former Soviet Union, China, and Mongolia are used for comparison with the diagnosed estimates of TWS variations. The datasets entail gravimetric measurements of the plant available soil water in different types of agricultural areas. An overview of the considered soil moisture datasets and of their temporal and vertical resolution is presented in Table 3. Snow depth measurements for the former Soviet Union spanning a period from 1881 to 1995 are also included in the comparisons. They are based on observations made by personnel at 284 WMO stations and are available on a CD-ROM (NSIDC 1999). The snow depths are converted in water equivalent using a snow density value of 100 kg m−2 (value for fresh snow, e.g., Brasnett 1999). Also larger values for the snow density representing older snow were tested. However, the resulting change in the results of the comparisons discussed below was only minor.
All available station measurements (see Fig. 1) are averaged over the corresponding basins assuming equal weight. Despite the known problems when averaging point measurements over an area, this procedure is applicable in regions with a reasonably homogeneous distribution of the stations and can give there an indication of the quality of the diagnosed TWS variations.
Besides soil moisture and snow water, TWS also includes mainly groundwater and surface water, which can significantly contribute to the total TWS variations (e.g., S04). In the regions investigated here, there are to our knowledge no comprehensive, large-scale observational datasets of groundwater and surface water. Therefore the comparison of the computed estimates with soil moisture and snow depth observations has only a reduced significance in regions where seasonal changes in groundwater and surface water play an important role in the water cycle. However, together with the findings of S04, they can give further evidence of the applicability of the water-balance method in different regions.
In Figs. 8 and 9, the diagnosed monthly variations in TWS (drift corrected) are compared against soil moisture and snow depth observations for the Volga, Ob, Dnepr, and Don River basins. Correlation between observed and diagnosed monthly variations amount to between 0.59 and 0.75. The agreement is also reasonable in terms of anomalies (i.e., deviation from the climatology) for these basins (not shown). In some other regions (e.g., Amur), the correlation is quite poor, likely due to the nonrepresentative distribution of the soil moisture and snow measurement stations in the basins (see Fig. 1) and due to the lack of groundwater and surface water data. The amplitudes of the observations often seem to be smaller than the estimates. Especially in summer, the decrease in water storage is less pronounced in the observations (e.g., Ob). The water-balance estimates also seem to slightly lag behind the observations. Both effects are likely associated with groundwater and surface water (wetlands, lakes, reservoirs, rivers) storage, which are not accounted for in the observations available. The west Siberian region, for example, is known to be one of the world largest wetlands (Fraser and Keddy 2005), and western and central Russia is an area of considerable groundwater recharge (up to 300 mm yr−1, estimated by the hydrological model WaterGAP Global Hydrological Model; Döll et al. 2003; Struckmeier et al. 2004). To account for the missing observations of temporarily stored groundwater and surface water in these comparisons, we also performed additional comparisons applying a time delay of one month for the runoff in the water-balance calculations for the large Ob and Volga River basins. As a consequence, the resulting correlations between the estimates and soil moisture and snow observations increased slightly (Ob: from 0.75 to 0.79; Volga: from 0.74 to 0.76), showing the importance of groundwater and surface water storage in these regions.
d. Comparison against ERA-40 soil moisture and snow
For comparison, simulated ERA-40 total TWS variations (see section 3a) are also included in the seasonal mean displays (see Figs. 2a,b and 3 –5). The ERA-40 data include soil moisture from all levels (0–7, 7–28, 28–100, and 100–289 cm), as well as snow water equivalent. As there is no groundwater in the ERA-40 model; soil moisture and snow represent the ERA-40 model's total TWS.
Compared to the computed estimates, the simulated ERA-40 TWS shows an earlier and stronger depletion in spring (March, April, May) for almost all investigated river basins, whereas its winter values are often higher than the estimated TWS variations (e.g., Odra, Dnepr, Mackenzie). The ERA-40 snow analysis (van den Hurk et al. 2000) uses snow depth observations, and in addition a nudging toward climatology in areas where observations are inadequate to correct the modeled snow depth. Analyses of the ERA-40 surface water budgets in the Mississippi (Betts et al. 2003a) and the Mackenzie (Betts et al. 2003b) River basins have shown that positive snow water analysis increments are comparable to snowfall. Being a significant source term in the frozen water budget, this addition of water in winter likely leads to the observed higher values of TWS variations in ERA-40. The early and strong decrease of TWS in ERA-40 is presumably associated with the fact that melted snow in the model is immediately removed as runoff, whereas in reality it might refreeze deeper in the snowpack (Betts et al. 2003b).
In summer, the ERA-40 TWS often shows a less pronounced decrease (e.g., Rhine, Ob, Mississippi). This behavior is likely related to the soil moisture nudging scheme used in ERA-40, which corrects soil moisture in the first three layers (0–7, 7–28, and 28–100 cm) based on the 6-h atmospheric analysis increments of specific humidity at the lowest model level (Douville et al. 2000). The idea behind this scheme is that a too dry (or too wet) soil will lead to a too dry (or too wet) boundary layer in the first-guess 6-h forecast compared to the atmospheric analysis (which includes information from humidity observations). The nudging is effective in preventing the model to drift. However, the resulting soil moisture increments during summer predominantly add water to the soil, leading to a damping of the seasonal cycle (Betts et al. 1998, 1999, 2003a; S04).
Overall, the diagnosed estimates of monthly TWS variations show a much more realistic seasonal cycle than the simulated ERA-40 values. Similar conclusions were reached in S04, showing in particular that the diagnosed water-balance estimates for Illinois agree better with in situ observations than the ERA-40 soil moisture.
5. Application to model validation
An important motivation of the current study is the use of diagnosed TWS variations for the validation of climate models. All current climate models contain some description of surface hydrology, but there is an apparent lack of validation data. For illustration of the envisioned approach, we here present some preliminary results on the validation of regional climate models over Europe.
The diagnosed estimates of monthly TWS variations in European river basins is used for the validation of regional climate models involved in the EU-project Prediction of Regional Scenarios and Uncertainties for Defining European Climate Change Risks and Effects (PRUDENCE; see http://prudence.dmi.dk/). The project is geared toward reducing deficiencies and quantifying uncertainties in projections of future climate change (Christensen et al. 2002). The models differ with respect to the physical and dynamical formulations, land-use characteristics, and computational domains (see Table 4). All models cover the major part of Europe at a resolution of approximately 50 km. They simulate a control climate (1961–90), as well as an A2-scenario time slice (2071–2100), and are driven by the Hadley Centre HadAM3H simulations [except the Centre National de Recherches Meteorologiques (CNRM) control run driven by observed sea surface temperatures]. Here attention is restricted to the control integration.
Figure 10 shows a comparison of the drift-corrected diagnosed monthly variations in TWS against control runs of various PRUDENCE models (see Table 4) and simulated ERA-40 TWS variations (see also section 4d) for the Danube basin. The total TWS of the models includes soil water and snow water equivalent. These first preliminary results display substantial differences between the models. Several models, as well as the ERA-40 reanalysis (as already mentioned in section 4d) underestimate the decrease in TWS during summer compared to the diagnosed water-balance estimates. The differences during summer are particularly large and amount to more than 1 mm day−1. Some models underestimate the summer drying by more than a factor 2. There are also considerable deficiencies in winter (likely relating to the representation of snow).
As a next step, a systematic analysis of the PRUDENCE models in the many European river basins will be conducted using the produced dataset of drift-corrected TWS variations. Another preliminary application can be found in van den Hurk et al. (2005; analysis of regional climate models in the Rhine basin). This illustrates the promising potential of the new dataset for model validation purposes.
6. Data download
The diagnosed basin-scale water-balance (BSWB) data of the various river basins described in this paper can be downloaded from the Web (at http://www.iac.ethz.ch/data/water_balance/).
7. Summary and conclusions
The combined terrestrial and atmospheric water-balance approach has been applied for the diagnosis of monthly terrestrial water storage variations in 37 river basins of Europe, Asia, North America, and Australia, following an earlier study over the Mississippi region (Seneviratne et al. 2004). The latter study showed excellent agreement with observational data of soil moisture, snow, and groundwater in a domain covering Illinois. The present paper extends this analysis to other regions of the midlatitudes and also performs some additional validation.
The long-term balance of the moisture flux convergence and the runoff shows negative imbalances in the computed water storage in many basins of Europe. Comparisons of the ERA-40 water vapor flux convergence with estimates from former water-balance studies for Europe point out the difficulty of averaging the ERA-40 grid fields over small areas affected by topography. A source of artificial drift is changes of the global observing system used in the 44 yr of the reanalysis. By applying a simple drift correction, these artificial imbalances can be corrected. Yet there remains some uncertainty as the drift correction may lead to errors if the underlying bias is associated with a strong seasonal cycle. The comparisons of the diagnosed terrestrial water storage variations against soil moisture observations in the former Soviet Union, Mongolia, and China, as well as with snow depth measurements in the former Soviet Union, display reasonably good correlations in basins with a representative distribution of soil moisture measurement stations (Volga, Ob, Don, and Dnepr), though the contribution of groundwater is generally difficult to assess on these large scales. The accuracy of the diagnosed water balances is contrasted for the various regions investigated and appears to be highest for large river basins with little topographic and climatic variations. Comparison with simulated ERA-40 terrestrial water storage variations showed that the diagnostic water balances have a much more realistic seasonal cycle.
The possible applications of the newly derived dataset are numerous, as has been shown by an exemplary application to model validation (section 5), as well as in the study of van den Hurk et al. (2005). It is expected that the new dataset (see previous section for data download) will find further climate applications in the future, in particular for inferring information on terrestrial water storage variations in regions without large-scale observations.
Acknowledgments
We thank Pedro Viterbo (ECMWF) for making available the ERA-40 data. Many thanks to the respective teams of the Global Soil Moisture Data Bank, the Global Runoff Data Centre (Koblenz, Germany; Thomas Lüllwitz) and the U.S. Geological Survey (USGS). We wish to thank also the local providers of runoff data (Michel Bouziges, Luigi Ciarmatori, Kevin Ellett, Simon Lery, Luc Levasseur, Frédéric Raout, Min Yawou; for affiliations see Table 1), as well as Joachim Gurtz, Massimiliano Zappa, and J. C. Scherer (Ministère de l'aménagement du territoire et de l'environnement, Bureau des données sur l'eau, Paris, France) for the help in collecting this data. Sincere thanks to Dietmar Grebner, Andreas Güntner, Bart van den Hurk, Simon Jaun, Thomas Peter, Reto Stöckli, and Pier Luigi Vidale for many useful comments and discussions, to Daniel Lüthi for computational support, and to Pia Eugster for the digitalization of runoff data. Comments from three anonymous reviewers are also gratefully acknowledged. This research was supported by the Fifth Framework Programme of the European Union (project PRUDENCE, Contract EVK2-2000-00132), by the Swiss Ministry for Education and Research (BBW Contract 01.0305-1), and by the Swiss National Science Foundation (NCCR Climate).
REFERENCES
Alestalo, M., 1983: The atmospheric water vapour budget over Europe. Variations in the Global Water Budget, A. Street-Perrott, Ed., D. Reidel, 67–79.
Baumgartner, A., and Reichel E. , 1975: The World Water Balance—Mean Annual Global, Continental and Maritime Precipitation, Evaporation and Runoff. Elsevier, 179 pp.
Beljaars, A. C. M., Viterbo P. , Miller M. J. , and Betts A. K. , 1996: The anomalous rainfall over the United States during July 1993: Sensitivity to land surface parameterization and soil moisture anomalies. Mon. Wea. Rev, 124 , 362–383.
Bengtsson, L., Hagemann S. , and Hodges K. I. , 2004a: Can climate trends be calculated from reanalysis data? J. Geophys. Res, 109 .D11111, doi:10.1029/2004JD004536.
Bengtsson, L., Hodges K. I. , and Hagemann S. , 2004b: Sensitivity of large-scale atmospheric analyses to humidity observations and its impact on the global water cycle and tropical and extratropical weather systems in ERA-40. Tellus, 56A , 202–217.
Berbery, E. H., and Rasmusson E. M. , 1999: Mississippi moisture budgets on regional scales. Mon. Wea. Rev, 127 , 2654–2673.
Betts, A. K., Ball J. H. , Beljaars A. C. M. , Miller M. J. , and Viterbo P. A. , 1996: The land–surface atmosphere interaction: A review based on observational and global modeling perspectives. J. Geophys. Res, 101 , 7209–7225.
Betts, A. K., Viterbo P. , and Wood E. , 1998: Surface energy and water balance for the Arkansas–Red River basin from the ECMWF reanalysis. J. Climate, 11 , 2881–2897.
Betts, A. K., Ball J. H. , and Viterbo P. , 1999: Basin-scale surface water and energy budgets for the Mississippi from the ECMWF reanalysis. J. Geophys. Res, 104 , 19293–19306.
Betts, A. K., Ball J. H. , Bosilovich M. , Viterbo P. , Zhang Y. , and Rossow W. B. , 2003a: Intercomparison of water and energy budgets for five Mississippi subbasins between ECMWF reanalysis (ERA-40) and NASA Data Assimilation Office fvGCM for 1990–1999. J. Geophys. Res, 108 .8618, doi:10.1029/2002JD003127.
Betts, A. K., Ball J. H. , and Viterbo P. , 2003b: Evaluation of the ERA-40 surface water budget and surface temperature for the Mackenzie River basin. J. Hydrometeor, 4 , 1194–1211.
Brasnett, B., 1999: A global analysis of snow depth for numerical weather prediction. J. Appl. Meteor, 38 , 726–740.
Budyko, M. I., Ed. 1963: Atlas of the Heat Balance of the Earth. Glavnaia Geofizica Obsercatoriia, 69 pp.
Christensen, J. H., Christensen O. B. , Lopez P. , van Meijgaard E. , and Botzet M. , 1996: The HIRHAM4 regional atmospheric climate model. Scientific Rep. 96-4, Danish Meteorological Institute, 51 pp.
Christensen, J. H., Carter T. R. , and Giorgi F. , 2002: PRUDENCE employs new methods to assess European climate change. Eos, Trans. Amer. Geophys. Union,83, 147.
Courtier, P., and Coauthors, 1998: The ECMWF implementation of three dimensional variational assimilation (3D-Var). Part I: Formulation. Quart. J. Roy. Meteor. Soc, 124 , 1783–1808.
Déqué, M., Marquet P. , and Jones R. G. , 1998: Simulation of climate change over Europe using a global variable resolution general circulation model. Climate Dyn, 14 , 173–189.
Dettinger, M. D., and Diaz H. F. , 2000: Global characteristics of stream flow seasonality and variability. J. Hydrometeor, 1 , 289–309.
Dirmeyer, P. A., Dolman A. J. , and Sato N. , 1999: The pilot phase of the Global Soil Wetness Project. Bull. Amer. Meteor. Soc, 80 , 851–875.
Dirmeyer, P. A., Gao X. , and Oki T. , 2002: GSWP-2: The Second Global Soil Wetness Project Science and Implementation Plan. IGPO Publication Series 37, International GEWEX Project Office, 65 pp.
Döll, P., Kaspar F. , and Lehner B. , 2003: A global hydrological model for deriving water availability indicators: Model tuning and validation. J. Hydrol, 270 , 105–134.
Döscher, R., Willen U. , Jones C. , Rutgersson A. , Meier H. E. M. , Hansson U. , and Graham L. P. , 2002: The development of the coupled regional ocean–atmosphere model RCAO. Boreal Environ. Res, 7 , 183–192.
Douville, H., Viterbo P. , Mahfouf J-F. , and Beljaars A. C. M. , 2000: Evaluation of the optimum interpolation and nudging techniques for soil moisture analysis using FIFE data. Mon. Wea. Rev, 128 , 1733–1756.
Eltahir, E. A. B., 1998: A soil moisture–rainfall feedback mechanism. 1. Theory and observations. Water Resour. Res, 34 , 765–776.
Entin, J. K., Robock A. , Vinnikov K. Y. , Zabelin V. , Liu S. , Namkhai A. , and Adysasuren T. , 1999: Evaluation of Global Soil Wetness Project soil moisture simulations. J. Meteor. Soc. Japan, 77 , 183–197.
Fraser, L. H., and Keddy P. A. , 2005: The World's Largest Wetlands: Ecology and Conservation. Cambridge University Press, 488 pp.
Frei, C., and Schär C. , 1998: A precipitation climatology of the Alps from high-resolution rain-gauge observations. Int. J. Climatol, 18 , 873–900.
Giorgi, F., Marinucci M. R. , and Bates G. T. , 1993a: Development of a second generation regional climate model (RegCM2). Part I: Boundary layer and radiative transfer processes. Mon. Wea. Rev, 121 , 2794–2813.
Giorgi, F., Marinucci M. R. , Bates G. T. , and Canio G. D. , 1993b: Development of a second generation regional climate model (REGCM2). Part II: Convective processes and assimilation of lateral boundary conditions. Mon. Wea. Rev, 121 , 2814–2832.
Gutowski, W. J. J., Chen Y. , and Ötles Z. , 1997: Atmospheric water vapor transport in NCEP–NCAR reanalyses: Comparison with river discharge in the central United States. Bull. Amer. Meteor. Soc, 78 , 1957–1969.
Hagemann, S., and Dümenil L. , 1998: A parametrization of the lateral waterflow for the global scale. Climate Dyn, 14 , 17–31.
Hortal, M., and Simmons A. J. , 1991: Use of reduced Gaussian grids in spectral models. Mon. Wea. Rev, 119 , 1057–1074.
Jacob, D., 2001: A note to the simulation of the annual and inter-annual variability of the water budget over the Baltic Sea drainage basin. Meteor. Atmos. Phys, 77 , 61–73.
Jones, R., Murphy J. , Hassell D. , and Taylor R. , 2001: Ensemble mean changes in a simulation of the European climate of 2071–2100, using the new Hadley Centre regional climate modelling system HadAM3H/HadRM3H. Hadley Centre Rep., Met Office, 19 pp. [Avaliable online at http://prudence.dmi.dk/public/publications/hadley_200208.pdf.].
Kleinn, J., Frei C. , Gurtz J. , Lüthi D. , Vidale P. L. , and Schär C. , 2005: Hydrologic simulations in the Rhine basin driven by a regional climate model. J. Geophys. Res, 110 .doi:10.1029/2004JD005143.
Korzun, V. I., 1974: World water balance and water resources of the earth. Report of the USSR Committee for the IHD, 663 pp.
Koster, R. D., Suarez M. J. , and Heiser M. , 2000: Variance and predictability of precipitation at seasonal-to-interannual timescales. J. Hydrometeor, 1 , 26–46.
Koster, R. D., and Coauthors, 2004: Regions of strong coupling between soil moisture and precipitation. Science, 305 , 1138–1140.
Lenderink, G., van den Hurk B. , van Meijgaard E. , van Ulden A. , and Cuijpers H. , 2003: Simulation of present-day climate in RACMO2: First results and model developments. KNMI Tech. Rep. 252, 24 pp.
Milly, P. C. D., and Dunne K. A. , 1994: Sensitivity of the global water cycle to the water-holding capacity of the land. J. Climate, 7 , 506–526.
Milly, P. C. D., Wetherald R. T. , Dunne K. A. , and Delworth T. L. , 2002: Increasing risks of great floods in a changing climate. Nature, 415 , 514–517.
Mintz, Y., and Serafini Y. V. , 1992: A global monthly climatology of soil moisture and water balance. Climate Dyn, 8 , 13–27.
NSIDC, 1999: Historical Soviet Daily Snow Depth. Version 2.0. National Snow and Ice Data Center, Boulder, CO, CD-ROM.
Oki, T., Musiake K. , Matsuyama J. , and Masuda K. , 1995: Global atmospheric water balance and runoff from large river basins. Hydrol. Processes, 9 , 655–678.
Pardé, M., 1947: Fleuves et Rivières. 2d ed. Librairie Armand Colin, 224 pp.
Peixoto, J. P., and Oort A. H. , 1992: Physics of Climate. AIP Press, 520 pp.
Pope, D. V., Gallani M. , Rowntree R. , and Stratton A. , 2000: The impact of new physical parameterizations in the Hadley Centre climate model: HadAM3. Climate Dyn, 16 , 123–146.
Rantz, S. E., 1982a: Measurement of stage and discharge. Vol. 1, Measurement and Computation of Streamflow, Water-supply Paper 2175, U.S. Geological Survey, 284 pp.
Rantz, S. E., 1982b: Computation of discharge. Vol. 2, Measurement and Computation of Streamflow, Water-supply Paper 2175, U.S. Geological Survey, 631 pp.
Rasmusson, E. M., 1968: Atmospheric water vapor transport and the water balance of North America. Mon. Wea. Rev, 96 , 720–734.
Reichle, R. H., Koster R. D. , Dong J. , and Berg A. A. , 2004: Global soil moisture from satellite observations, land surface models, and ground data: Implications for data assimilation. J. Hydrometeor, 5 , 430–442.
Robock, A., Vinnikov K. Y. , Srinivasan G. , Entin J. K. , Hollinger S. E. , Speranskaya N. A. , Liu S. , and Namkhai A. , 2000: The Global Soil Moisture Data Bank. Bull. Amer. Meteor. Soc, 81 , 1281–1299.
Robock, A., Mu M. , Vinnikov K. , Trofimova I. V. , and Adamenko T. I. , 2005: Forty five years of observed soil moisture in the Ukraine: No summer desiccation (yet). Geophys. Res. Lett, 32 .doi:10.1029/2004GL021914.
Rodell, M., and Famiglietti J. S. , 2001: An analysis of terrestrial water storage variations in Illinois with implications for the Gravity Recovery and Climate Experiment (GRACE). Water Resour. Res, 37 , 1327–1339.
Sanchez, E., Gallardo C. , Gaertner M. A. , Arribas A. , and Castro M. , 2004: Future climate extreme events in the Mediterranean simulated by a regional climate model: First approach. Global Planet. Change, 44 , 163–180.
Schär, C., Lüthi D. , Beyerle U. , and Heise E. , 1999: The soil–precipitation feedback: A process study with a regional climate model. J. Climate, 12 , 722–741.
Schär, C., Vasilina L. , Pertziger F. , and Dirren S. , 2004a: Seasonal runoff forecasting using precipitation from meteorological data assimilation systems. J. Hydrometeor, 5 , 959–973.
Schär, C., Vidale P. L. , Lüthi D. , Frei C. , Häberli C. , Liniger M. A. , and Appenzeller C. , 2004b: The role of increasing temperature variability in European summer heatwaves. Nature, 427 , 332–336.
Seneviratne, S. I., 2003: Terrestrial water storage: A critical variable for midlatitude climate and climate change. Ph.D. thesis, Atmospheric and Climate Science ETH Zürich, ETH No. 14944, 155 pp.
Seneviratne, S. I., Pal J. S. , Eltahir E. A. B. , and Schär C. , 2002: Summer dryness in a warmer climate: A process study with a regional climate model. Climate Dyn, 20 , 69–85.
Seneviratne, S. I., Viterbo P. , Lüthi D. , and Schär C. , 2004: Inferring changes in terrestrial water storage using ERA-40 reanalysis data: The Mississippi River basin. J. Climate, 17 , 2039–2057.
Shukla, J., and Mintz Y. , 1982: Influence of land-surface evaporation on the earth's climate. Science, 215 , 1498–1501.
Simmons, A. J., and Gibson J. K. , 2000: The ERA-40 Project Plan. ERA-40 Project Report Series 1, ECMWF, Shinfield Park, Reading, United Kingdom, 62 pp.
Simmons, A. J., and Coauthors, 2004: Comparison of trends and low-frequency variability in CRU, ERA-40 and NCEP/NCAR analyses of surface air temperature. J. Geophys. Res, 109 .D24115, doi:10.1029/2004JD005306.
Steppeler, J., Doms G. , Schättler U. , Bitzer H. W. , Gassmann A. , Damrath U. , and Gregoric G. , 2003: Meso-gamma scale forecasts using the nonhydrostatic model LM. Meteor. Atmos. Phys, 82 , 75–96.
Struckmeier, W. F., Gilbrich W. H. , Richts A. , and Zaepke M. , 2004: WHYMAP and the groundwater resources map of the world at the scale of 1:50 000 000: Special Edition for the 32nd International Geological Congress, Florence/Italy, 20–28 August 2004. BGR. [Available online at http://www.bgr.de/b1hydro/fachbeitraege/a200401/expl_note.pdf.].
Tapley, B. D., Bettadpur S. , Ries J. C. , Thompson P. F. , and Watkins M. M. , 2004: GRACE measurements of mass variability in the earth system. Science, 305 , 503–505.
van den Hurk, B. J. J. M., Viterbo P. , Beljaars A. C. M. , and Betts A. K. , 2000: Offline validation of the ERA40 surface scheme. Tech. Memo. 295, ECMWF, Reading, United Kingdom, 43 pp.
van den Hurk, B. J. J. M., and Coauthors, 2005: Soil control on runoff response to climate change in regional climate model simulations. J. Climate, 18 , 3536–3551.
Vidale, P. L., Lüthi D. , Frei C. , Seneviratne S. , and Schär C. , 2003: Predictability and uncertainty in a regional climate model. J. Geophys. Res, 108 .4586, doi:10.1029/2002JD002810.
Vörösmarty, C. J., Fekete B. M. , Meybeck M. , and Lammers R. , 2000: A simulated topological network representing the global system of rivers at 30-minute spatial resolution (STN-30). Global Biogeochem. Cycles, 14 , 599–621.
Wahr, J., Swenson S. , Zlotnicki V. , and Velicogna I. , 2004: Time-variable gravity from GRACE: First results. Geophys. Res. Lett, 31 .L11501, doi:10.1029/2004GL019779.
Wetherald, R. T., and Manabe S. , 1999: Detectability of summer dryness caused by greenhouse warming. Climatic Change, 43 , 495–511.
Winter, T. C., 1981: Uncertainties in estimating the water balance of lakes. Water Resour. Bull, 17 , 82–115.
Yeh, P. J-F., Irizarry M. , and Eltahir E. A. B. , 1998: Hydroclimatology of Illinois: A comparison of monthly evaporation estimates based on atmospheric water balance and soil water balance. J. Geophys. Res, 103 , 19823–19837.
Zektser, I. S., and Loaiciga H. A. , 1993: Groundwater fluxes in the global hydrologic cycle: Past, present and future. J. Hydrol, 144 , 405–427.
Zektser, I. S., Loaiciga H. A. , and Wolf J. T. , 2005: Environmental impacts of groundwater overdraft: Selected case studies in the southwestern United States. Environ. Geol, 47 , 396–404.
Investigated river basins in (a) Europe (vertically hatched: central European basins; horizontally hatched: French basins), (b) Asia, (c) North America (horizontally hatched: whole Mississippi basin), and (d) Australia, as well as soil moisture measurement stations (×) and snow observations (▪).
Citation: Journal of Hydrometeorology 7, 1; 10.1175/JHM480.1
Investigated river basins in (a) Europe (vertically hatched: central European basins; horizontally hatched: French basins), (b) Asia, (c) North America (horizontally hatched: whole Mississippi basin), and (d) Australia, as well as soil moisture measurement stations (×) and snow observations (▪).
Citation: Journal of Hydrometeorology 7, 1; 10.1175/JHM480.1
Investigated river basins in (a) Europe (vertically hatched: central European basins; horizontally hatched: French basins), (b) Asia, (c) North America (horizontally hatched: whole Mississippi basin), and (d) Australia, as well as soil moisture measurement stations (×) and snow observations (▪).
Citation: Journal of Hydrometeorology 7, 1; 10.1175/JHM480.1
(a), (b) Mean seasonal cycles (mm day−1) of water vapor flux convergence (denoted −∇Q), surface runoff, simulated ERA-40 total TWS variations, and diagnosed variations in TWS [denoted ΔS, raw and drift-corrected (corr) estimates] for the (left) Ob and (right) Rhine River basins. The long-term means of the uncorrected diagnosed TWS variations [denoted ΔS (avg)] are listed as well. (c), (d) Cumulated water vapor flux convergence, runoff, and diagnosed TWS S (mm) for the same basins (uncorrected and drift-corrected estimates, starting value set to 0 mm). In (d), the uncorrected TWS is also shown for the Rhine basin downstream of Basel.
Citation: Journal of Hydrometeorology 7, 1; 10.1175/JHM480.1
(a), (b) Mean seasonal cycles (mm day−1) of water vapor flux convergence (denoted −∇Q), surface runoff, simulated ERA-40 total TWS variations, and diagnosed variations in TWS [denoted ΔS, raw and drift-corrected (corr) estimates] for the (left) Ob and (right) Rhine River basins. The long-term means of the uncorrected diagnosed TWS variations [denoted ΔS (avg)] are listed as well. (c), (d) Cumulated water vapor flux convergence, runoff, and diagnosed TWS S (mm) for the same basins (uncorrected and drift-corrected estimates, starting value set to 0 mm). In (d), the uncorrected TWS is also shown for the Rhine basin downstream of Basel.
Citation: Journal of Hydrometeorology 7, 1; 10.1175/JHM480.1
(a), (b) Mean seasonal cycles (mm day−1) of water vapor flux convergence (denoted −∇Q), surface runoff, simulated ERA-40 total TWS variations, and diagnosed variations in TWS [denoted ΔS, raw and drift-corrected (corr) estimates] for the (left) Ob and (right) Rhine River basins. The long-term means of the uncorrected diagnosed TWS variations [denoted ΔS (avg)] are listed as well. (c), (d) Cumulated water vapor flux convergence, runoff, and diagnosed TWS S (mm) for the same basins (uncorrected and drift-corrected estimates, starting value set to 0 mm). In (d), the uncorrected TWS is also shown for the Rhine basin downstream of Basel.
Citation: Journal of Hydrometeorology 7, 1; 10.1175/JHM480.1
As in Figs. 2a and 2b but for other European basins.
Citation: Journal of Hydrometeorology 7, 1; 10.1175/JHM480.1
As in Figs. 2a and 2b but for other European basins.
Citation: Journal of Hydrometeorology 7, 1; 10.1175/JHM480.1
As in Figs. 2a and 2b but for other European basins.
Citation: Journal of Hydrometeorology 7, 1; 10.1175/JHM480.1
As in Figs. 2a and 2b but for basins in Russia and Asia.
Citation: Journal of Hydrometeorology 7, 1; 10.1175/JHM480.1
As in Figs. 2a and 2b but for basins in Russia and Asia.
Citation: Journal of Hydrometeorology 7, 1; 10.1175/JHM480.1
As in Figs. 2a and 2b but for basins in Russia and Asia.
Citation: Journal of Hydrometeorology 7, 1; 10.1175/JHM480.1
As in Figs. 2a and 2b but for basins in North America and Australia.
Citation: Journal of Hydrometeorology 7, 1; 10.1175/JHM480.1
As in Figs. 2a and 2b but for basins in North America and Australia.
Citation: Journal of Hydrometeorology 7, 1; 10.1175/JHM480.1
As in Figs. 2a and 2b but for basins in North America and Australia.
Citation: Journal of Hydrometeorology 7, 1; 10.1175/JHM480.1
Regions A (vertically hatched) and M (horizontally hatched) from Alestalo (1983), used for comparisons with ERA-40 water vapor convergence.
Citation: Journal of Hydrometeorology 7, 1; 10.1175/JHM480.1
Regions A (vertically hatched) and M (horizontally hatched) from Alestalo (1983), used for comparisons with ERA-40 water vapor convergence.
Citation: Journal of Hydrometeorology 7, 1; 10.1175/JHM480.1
Regions A (vertically hatched) and M (horizontally hatched) from Alestalo (1983), used for comparisons with ERA-40 water vapor convergence.
Citation: Journal of Hydrometeorology 7, 1; 10.1175/JHM480.1
Long-term imbalances of water storage {
Citation: Journal of Hydrometeorology 7, 1; 10.1175/JHM480.1
Long-term imbalances of water storage {
Citation: Journal of Hydrometeorology 7, 1; 10.1175/JHM480.1
Long-term imbalances of water storage {
Citation: Journal of Hydrometeorology 7, 1; 10.1175/JHM480.1
Comparison of diagnosed monthly TWS variations (drift corrected) against combined in situ observations of soil moisture and snow depth variations (mm day−1) in the (a) Volga and (b) Ob River basins. (top) Time series, (middle) climatologies, and (bottom) scatterplots with monthly resolution. Note that groundwater is not included due to the lack of observed data.
Citation: Journal of Hydrometeorology 7, 1; 10.1175/JHM480.1
Comparison of diagnosed monthly TWS variations (drift corrected) against combined in situ observations of soil moisture and snow depth variations (mm day−1) in the (a) Volga and (b) Ob River basins. (top) Time series, (middle) climatologies, and (bottom) scatterplots with monthly resolution. Note that groundwater is not included due to the lack of observed data.
Citation: Journal of Hydrometeorology 7, 1; 10.1175/JHM480.1
Comparison of diagnosed monthly TWS variations (drift corrected) against combined in situ observations of soil moisture and snow depth variations (mm day−1) in the (a) Volga and (b) Ob River basins. (top) Time series, (middle) climatologies, and (bottom) scatterplots with monthly resolution. Note that groundwater is not included due to the lack of observed data.
Citation: Journal of Hydrometeorology 7, 1; 10.1175/JHM480.1
As in Fig. 8 but for the (a) Dnepr and (b) Don basins.
Citation: Journal of Hydrometeorology 7, 1; 10.1175/JHM480.1
As in Fig. 8 but for the (a) Dnepr and (b) Don basins.
Citation: Journal of Hydrometeorology 7, 1; 10.1175/JHM480.1
As in Fig. 8 but for the (a) Dnepr and (b) Don basins.
Citation: Journal of Hydrometeorology 7, 1; 10.1175/JHM480.1
Comparison of drift-corrected estimates of variations in TWS against PRUDENCE model runs and simulated ERA-40 data (mm day−1) for the Danube basin (772 220 km2). The diagnosed water-balance estimates are based on the period 1961–90, and the simulations are representative for the same period but employ the forcing of an AGCM run (HadAM3H) driven by observed sea surface temperature variations.
Citation: Journal of Hydrometeorology 7, 1; 10.1175/JHM480.1
Comparison of drift-corrected estimates of variations in TWS against PRUDENCE model runs and simulated ERA-40 data (mm day−1) for the Danube basin (772 220 km2). The diagnosed water-balance estimates are based on the period 1961–90, and the simulations are representative for the same period but employ the forcing of an AGCM run (HadAM3H) driven by observed sea surface temperature variations.
Citation: Journal of Hydrometeorology 7, 1; 10.1175/JHM480.1
Comparison of drift-corrected estimates of variations in TWS against PRUDENCE model runs and simulated ERA-40 data (mm day−1) for the Danube basin (772 220 km2). The diagnosed water-balance estimates are based on the period 1961–90, and the simulations are representative for the same period but employ the forcing of an AGCM run (HadAM3H) driven by observed sea surface temperature variations.
Citation: Journal of Hydrometeorology 7, 1; 10.1175/JHM480.1
Characteristics of the analyzed river basins.
Comparison of annual mean values of − {
Characteristics of the employed soil moisture datasets (Global Soil Moisture Data Bank; Robock et al. 2000).
Description of the PRUDENCE models (see http://prudence.dmi.dk/private/computing/model_descrip.html). The acronyms in parentheses refer to Fig. 10.
