WATCH Technical Report Number 44: Impact of spatial and temporal resolution on modelled terrestrial hydrological cycle components The impact of both spatial and temporal resolution on the components of the terrestrial hydrological cycle are investigated using the WATCH forcing dataset (WFD) and the JULES (Joint UK Land Environment Simulator) land surface model. The various spatial resolutions are achieved by degrading the native half degree latitude/longitude resolution WATCH dataset to both one degree and two degrees. The temporal resolutions are created by degrading the native three hourly WATCH forcing dataset to six hourly and using the WATCH interpolator to derive a one hour forcing dataset. There is little difference in the moisture stores of soil water and canopy water in the long term mean from the various resolutions, the presented analysis is for changes in evaporation and runoff. Evaporation is further analysed into its various components for the spatial resolution. Results suggest little impact from spatial resolution, but the interpolation method for temporal resolution can have a significant effect on total mean evaporation/runoff balance.
Emma Compton and Martin Best, July 22, 2011
WATCH Technical Report Number 43: Drought at the global scale in the 21st Century This study has explored differences in drought in total runoff (surface runoff and subsurface runoff) at the global and the continental scale between seven different large-scale models (global hydrological models and land surface models). First drought characteristics (number of droughts and average drought duration) were investigated for the control period (1971-2000). Droughts derived from the large-scale models forced with WATCH Forcing Data (WFD, re-analysis dataset) and downscaled, bias-corrected climate output from three different GCMs (CNRM, ECHAM and IPSL) were intercompared. Next droughts for the 21st century (2021-2050 and 2071-2100) were explored to investigate different climate change, i.e. drought characteristics derived from large-scale models that were forced with climate output from the three GCMs for two emission scenarios (A2 and B1). The number and spatial distribution of drought events, and average drought duration do not clearly show a consistent change (increase or decrease) among the models due to variation of number of arid cells (cells with zero flow) that are found in the different simulation periods (control, mid and late 21st century). After the introduction of an equivalent drought measure, to remove arid cells, it seems that the large-scale models have relatively similar statistical values and therefore similar statistical properties for the WFD and GCM climate output. An increase in number of drought events was identified over the 21st century.
Gerald A. Corzo Perez, Henny A.J. van Lanen, Nathalie Bertrand, Cui Chen, Douglas Clark, Sonja Folwell, Simon N. Gosling, Naota Hanasaki, Jens Heinke & Frank Voβ, August 25, 2011
WATCH Technical Report Number 42: Drought at the global scale for the 2nd part of the 20th Century (1963-2001) Simulated gridded time series from ten different large-scale models, both GHMs and LSMs, are used to investigate to what extent these large-scale models capture historic hydrological drought events. Monthly aggregated total runoff values were used to determine hydrological drought events with different methods developed to identify spatio-temporal drought characteristics. For each continent one documented major drought event was selected and described in detail. It can be concluded that most major drought events are captured by the models, but the spatial extent of the drought events differ substantially between the models.
Van Huijgevoort, M.H.J., Hazenberg, P. van Lanen, H.A.J., Bertrand, N., Clark, D., Folwell, S., Gomes, S., Gosling, S., Hanasaki, N., Heinke, J. & Koirala, S., Stacke, T. & Voβ, F., July 27, 2011
WATCH Technical Report Number 41: High flow in the 21st Century: analysis with a simple conceptual hydrological models using the input of 3 GCMs (A2 scenario) In this study a conceptual hydrological model was applied, which was forced with the outcome from three GCMs (ECHAM, CNRM, IPSL) for the A2 scenario in the 21st century. The Q10 (streamflow that is equalled or exceeded in 10% of the time) was used a high flow metric. It was calculated for 1495 selected land points that are well distributed across the globe, which adequately represent the Köppen-Geiger major climate regions. The hydrological model was run for a soil with an intermediate soil water supply capacity and an intermediary responding groundwater system. First, high flows obtained with the hydrological model using the three GCMs were checked against those got from the re-analysis dataset (WATCH Forcing Data) for the control period (CTRL, 1970-2000). The median Q10s derived from the GCMs for the CTRL period differ at least tens of per cent from the Q10 obtained when using WFD as input for the hydrological model (i.e. about 30-60%). The differences for the arid and polar (B- and E-) climates are even higher (about 100% or more). The impact of climate change (median of Q10) is projected to increase over the 21st century according to all three GCMs for the C-, D- and E-climates. This is also expected for the A- and B-climates according to ECHAM and CNRM. IPSL provides for these two Köppen-Geiger major climate regions a more diffuse projection.
Van Lanen, H.A.J. & Wanders, N., August 25, 2011
WATCH Technical Report Number 40: Multi-model analysis of drought at the global scale: differences in hydrological drought between the first and the second part of the 20th Century Global hydrological drought for the twentieth century has been investigated based on the simulated total runoff (sum of the surface and sub-surface runoff) of two global hydrological models (WaterGAP and GWAVA), two land surface models (HTESSEL and Orchidee), and an ensemble mean derived from six global models. Global hydrological drought trend (two-tailed t-test, Mann-Kendall non parametric test) was analysed by splitting the 20th century in to the first part (1906-1957) and the second part (1958-2000) based on the different approaches used to generate the WATCH meteorological forcing data. Trend shifts in simulated total runoff between the two periods of the century are observed in nearly half of the land surface cells. Sliding time windows were used to determine thresholds to smooth the effect of these trend shifts and also to capture inter-decadal climatic variability. Percentage of the globe and the continents in drought was computed for the whole simulation period at a monthly time step for each model. Most of the world wide droughts, that are often associated with the ENSO phenomena, are identified by the four models as well as the multi-model ensemble mean though there exist differences in percentage of area in drought among models. Inter-comparing the two periods of the century, decreasing trends are observed in area of the globe in drought for the first part (1906 -1958) of the century, whereas increasing trends are observed for the second part(1958-2000) of the century among all models.
Estifanos, S.T., van Huijgevoort, M.H.J., Hazenberg, P., van Lanen, H.A.J. & Weedon, G.P., August 25, 2011
WATCH Technical Report Number 39: Reference evapotranspiration with radiation-based and temperature-based method - impact on hydrological drought using WATCH Forcing Data Two different methods to calculate reference evapotranspiration were applied and resulting drought characteristics across the globe were compared. WATCH Forcing Data (WFD) have been used as input. One method is the so-called radiation based (ET0rad), and is the well-known Penman-Monteith equation with, among others, incoming short wave radiation as a variable. The other method is so-called temperature-based (ET0temp), the radiation term is replaced by an approximation of the radiation based on minimum and maximum air temperature. After calculating reference evapotranspiration ET0 with both methods, it is fed into a conceptual hydrological model that combines a soil water balance and a simple lumped groundwater model. The model generates daily actual evapotranspiration (ETA), groundwater recharge and discharge. Eventually, the simulated groundwater discharge was used to define periods of drought with a Variable Threshold (VT) method. Drought characteristics, like length and deficit, were identified. The model using the two different ET0 series, was applied to about 1500 cells (land grids) that well represent the five Koeppen-Geiger major climates across the world. There are clear differences between the two methods, and generally the radiation-based method leads to higher ET0 than the temperature-based method. As expected, the differences between ETArad and ETAtemp are significantly smaller than for the reference evapotranspiration. In general the radiation-based method still leads to higher ETA than the temperature-based method. Eventually, the different drought characteristics are compared. There are remarkable differences in drought characteristics, especially in deficit volume and intensity, but in general the differences are within ranges found in other literature that describes the impact of hydrological models or datasets with different diurnal forces on hydrological drought. The D and E-climates (snow-affected climates) show the largest differences.
Melsen, L.A., van Lanen, H.A.J. Wanders, N., van Huijgevoort, M.H.J. & Weedon, G.P., July 27, 2011
WATCH Technical Report Number 38: Including climate feedbacks in regional water resource assessments Two important changes to the meteorological conditions can occur as a result of a change in the land-surface state: the precipitation can change (Koster et al, 2004, Wang et al, 2007) and the evaporative demand can change (Ek and Holtslag, 2004, Schubert et al, 2004). Precipitation is particularly difficult to predict as it can be affected by large-scale weather patterns or complex processes such as mesoscale circulations or convective processes. The physics involved and the scale and complexity of the processes means that complex numerical atmospheric models are usually necessary to quantify impact of the land surface on precipitation, although a simple, analytical model can sometimes be used to assess the effect of the land- surface state on the likelihood of triggering convective precipitation (e.g., De Ridder, 1997). Changes in the evaporative demand are equally important. For instance, Cai et al (2009) have demonstrated the role that the land-atmosphere feedbacks have had on the recent Australian drought.
The feedback link between the land surface and the evaporative demand is through the Planetary Boundary Layer (PBL) and this feedback process lends itself to a fairly simple analysis. A framework to assess the impact of change in the land surface on the evaporative demand, and subsequent available regional water resources is proposed in this report.
Eleanor Blyth and Cor Jacobs, July 20, 2011
WATCH Technical Report Number 37: Creation of the WATCH 20th Century Ensemble Product One of the goals for WATCH was to assess the global terrestrial water cycle in the twentieth century. Here the creation of a new data set is described which provides a "best estimate" of four key hydrological variables based on averages of daily output from seven hydrological models. These models were operated on naturalised conditions (i.e. excluding human influences such as dams and irrigation). The outputs from these models resulted in full twentieth century runs, which provided a 100 years of data for the full twentieth century. The WATCH 20th Century Ensemble contains daily averages and associated descriptors for every half-degree land grid box in the Watch Forcing Data and are stored in monthly full latitude-longitude grid netCDF files. This report describe creation and availability of this ensemble product.
Graham P. Weedon et al, July 27, 2011
WATCH Technical Report Number 36: Changes in land surface model outputs due to uncertainty in the WATCH Forcing Data Uncertainty in the terrestrial water cycle, as represented by land surface model outputs, is investigated as a function of uncertainty in the WATCH Forcing Data. Three alternative forcing datasets were created involving changes to: a) rainfall and snowfall, b) downward shortwave and c) downward longwave and downward shortwave-radiation fluxes. The JULES land surface model was run using the standard and alternative forcing data to investigate changes in global snow water equivalent, evaporation, soil moisture and runoff.
Uncertainty in model outputs due to the forcing data is far smaller than uncertainty due to choice of hydrological model for all four hydrological variables. For snow water equivalent there is a very sensitive dependence on changes in precipitation and incoming radiation drivers such that relative changes in output represent an amplification of the relative changes in the forcing data input. For other variables the relative changes in outputs are in approximate proportion to relative changes in the input, except in semi-arid and arid regions where there is amplification for evaporation and runoff.
Graham P. Weedon, Martin Best, Douglas B. Clark, July 27, 2011
WATCH Technical Report Number 35: Assessment of flood peak simulations by Global Hydrological Models With significant changes to flood frequency anticipated as a result of climate change it becomes important to investigate how global hydrological models process climate forcing data. This study has analysed the variability in extreme precipitation and flood events from three global hydrological models driven by Watch Forcing Data and Echam5 climate model data, in order to assess variation in extreme rainfall-runoff processes, model bias, and the impact of climate change on the probabilistic behaviour of floods. Global spatial patterns of variation in extreme rainfall-runoff processes differ between models, particularly in regions of extreme climate. Flood statistics from hydrological models run with the same forcing data differ significantly. Climate change simulations indicate some general agreement between models in the emerging spatial pattern of future changes to flood variability across Europe; however, some distinct regional and sub-regional differences in magnitude of change and spatial pattern are evident between models.
James Miller, Thomas Kjeldsen, Christel Prudhomme, July 27, 2011
WATCH Technical Report Number 34: Improved hydrological processes in land surface models During the 4 years of WATCH the participating land-surface models progressed in their description of the continental water cycle. In particular 5 models participating in WATCH have been developed further in the course of the project. The main focus of the development of these models were the impact of human activities on the continental water budget. It was the comparison with global hydrological models during WATCH that lead land surface modelling groups to decide that the anthropogenic pressure on the water cycle was a major element missing in their models. Within WATCH some of the land-surface models introduced, for the first time, the impact of human activities on the continental water cycle. Model development and inclusion of anthropogenic influences will allow, in the future, to predict the evolution of the temperature of water bodies. An impact of climate change which will have important consequences for the ecology and industrial use of water bodies.
Jan Polcher, Nathalie Bertrand, Hester Biemans, Douglas B. Clark, Marina Floerke, Nicola Gedney, Dieter Gerten, Tobias Stacke, Michelle Van Vliet, Frank Voss, July 27, 2011
WATCH Technical Report Number 33: Objective low flow catalogue for Europe Regional drought and high flow characteristics vary spatially and temporally throughout Europe, and therefore objective classifications of drought and high flows are preferred in determining the variable regional characteristics of hydrological extremes. This study applies objective methodologies for defining regional droughts (Regional Deficiency Index; RDI) and high flow episodes (Regional High Flow Index; RHFI) to observed daily river flow timeseries from 579 stations across Europe, over the period 1961-2005. Drought and high flow catalogues are presented for 23 homogeneous European regions, allowing an assessment of the regional variation in the extent, duration, seasonality and spatio-temporal evolution of hydrological extremes in Europe.
Simon Parry, Christel Prudhomme, Jamie Hannaford, Jennifer Williamson, August 24, 2011