The hydroclimatic heterogeneity of the ColoradoRiverBasin complicates understanding its hydrology and hydro- logic response to projected changes in climate. The latitude of the water-producing regions in the basin lies at the north- ern boundary of the area in the American Southwest in which earlier studies have projected declines in precipitation and runoff (Milly et al., 2005; Seager et al., 2007; Seager and Vecchi, 2010). To the north of this boundary, in contrast, studies have tended to predict increases in rainfall and runoff. If the future boundary instead falls south of the ColoradoRiverBasin headwaters (water producing regions), stream- flows in the UpperColoradoRiverBasin will be reduced less than projected, or may increase. Given the large uncertainty over futureclimate evolution at the scale of this transition zone, advances in climate science (perhaps including higher- resolution earth system models) will be required before these consequences can be projected with much confidence (Sea- ger and Vecchi, 2010). The topography of the basin also in- fluences the interaction between climate and hydrology – the majority of precipitation in the basin falls in a very small fraction of its area (Reclamation, 2011b), mostly in the form of snow that is stored over seasonal time scales. Snow ac- cumulation and ablation are the most significant processes affecting the timing of streamflow generation in the basin, and the dynamics of the snow cycle will be significantly im- pacted by rising temperatures regardless of changes in pre- cipitation. Our comprehension of variations in snow abla- tion is currently grappling with quantifying uncertainties due to sporadic deposition of dust on the surface of the snow- pack, which accelerates snowmelt and affects basin runoff efficiency (Painter et al., 2007, 2010). The diversity of topog- raphy in the basin is another complicating factor. Outside of the cold, high-elevation areas, evapotranspiration is the dom- inant process affecting the water budget. Basin efficiencies (outflows divided by precipitation) range from about 30 % for sub-basins at higher average elevations to virtually zero in lower-elevation sub-basins (Reclamation, 2011b). The av- erage runoff efficiency of the entire basin is approximately 15 %. These factors, among others, contribute to the diffi- culty of quantifying and simulating the hydrology and hy- drologic sensitivity of the basin to climate forcings.
evaluated by use of a regional climate model (RCM) coupled with a hydrologic model, Soil and Water Assessment Tool (SWAT). The RCM we used resolves, at least partially, some fine-scale dynamical processes that are important contributors to precipitation in this region and that are not well simulated by global models. The SWAT model was calibrated and validated against measured streamflow data using observed weather data and inputs from the U.S. Environmental Protection Agency Better Assessment Science Integrating Point and Nonpoint Sources (BASINS) geographic information systems/ database system. Combined performance of SWAT and RCM was examined using observed weather data as lateral boundary conditions in the RCM. The SWAT and RCM performed well, especially on an annual basis. Potential impacts of climatechange on water yield and other hydrologic budget components were then quantified by driving SWAT with current and futurescenario climates. Twenty-one percent increase in future precipitation simulated by the RCM produced 18% increase in snowfall, 51% increase in surface runoff, and 43% increase in groundwater recharge, resulting in 50% net increase in total water yield in the Upper Mississippi RiverBasin on an annual basis. Uncertainty analysis showed that the simulated change in streamflow substantially exceeded model biases of the combined modeling system (with largest bias of 18%). While this does not necessarily give us high confidence in the actual climatechange that will occur, it does demonstrate that the climatechange ‘‘signal’’ stands out from the climate modeling (global plus regional) and impact assessment modeling (SWAT) ‘‘noise.’’ I NDEX T ERMS : 1655 Global Change: Water cycles (1836); 1860 Hydrology: Runoff and streamflow; 1866 Hydrology: Soil moisture; K EYWORDS : climatechange, streamflow, SWAT
Water managers have traditionally relied on the assump- tion of hydroclimatic stationarity to efficiently manage wa- ter resources and environmental operations. The timing and magnitude of runoff events is of particular importance, as ac- tual and forecasted runoff events can impact the operation of reservoirs (e.g., release schedules and magnitudes); how- ever, climatechange and anthropogenic alterations to basin characteristics increase the difficulty in accurately projecting streamflow conditions within hydrologic systems (e.g., Vil- larini et al., 2009). Raff et al. (2009) developed a method- ology to assess flood risk and runoff projections using pro- jections of futureclimate. Raff et al. (2009) utilized temper- ature and precipitation data from 112 Global Climate Mod- els (GCMs) within the World Climate Research Programme (WCRP) Coupled Model Intercomparison Project phase 3 (CMIP3) multi-model dataset (Meehl et al., 2007) subjected to statistical downscaling and bias-correction (Maurer et al., 2007) to drive the National Weather Service (NWS) River Forecasting System (RFS) hydrologic model. Each of the four basins investigated in Raff et al. (2009) exhibited the potential for increased flood frequency under changing cli- mate conditions, although the authors did acknowledge the need for further study to more fully understand these results. Other recent studies have developed alternative method- ologies for incorporating temperature and precipitation pat- terns over the UpperColoradoRiverBasin (Matter et al., 2010). Christensen and Lettenmaier (2007) has previously
There is increasing evidence that the interannual (year to year) and longer scale variations in the western U.S. hydro-climate are driven by large scale climate features such as the El Niño Southern Oscillation (ENSO), Pacific Decadal Oscillation (PDO), Atlantic Multi-decadal Oscillation (AMO), etc. [Hunter et al., 2006; McCabe and Dettinger, 1999; Piechota et al., 1997; Tootle et al., 2005]. Presently, stochastic simulation methods for the purpose of planning and water management aim to capture the distributional statistics (mean, variance, skew, etc.) of the observed record, but do not explicitly reflect the links between large scale, persistent climate features and streamflow. Furthermore, studies pertaining to simulation/projection of future hydrologic conditions often focus on the inter-annual time scale (1-2 years) or at the climatechange scale (50-100 years), leaving a substantial gap in the decadal to inter-decadal time scales. However, much of water management and planning focuses on meeting demands and reducing risk over decadal time scales. Thus, there is a clear need for tools that can generate stochastic traces and hydrologic projections at this time scale.
ABSTRACT: The Soil and Water Assessment Tool (SWAT) model was used to assess the effects of potential futureclimatechange on the hydrology of the Upper Mississippi RiverBasin (UMRB). Calibration and validation of SWAT were performed using monthly stream flows for 1968-1987 and 1988-1997, respectively. The R 2 and Nash-Sutcliffe simulation efficiency values computed for the monthly comparisons were 0.74 and 0.69 for the calibration period and 0.82 and 0.81 for the valida- tion period. The effects of nine 30-year (1968 to 1997) sensitiv- ity runs and six climatechange scenarios were then analyzed, relative to a scenario baseline. A doubling of atmospheric CO 2 to 660 ppmv (while holding other climate variables constant) resulted in a 36 percent increase in average annual streamflow while average annual flow changes of -49, -26, 28, and 58 per- cent were predicted for precipitation change scenarios of -20, -10, 10, and 20 percent, respectively. Mean annual streamflow changes of 51, 10, 2, -6, 38, and 27 percent were predicted by SWAT in response to climatechange projections generated from the CISRO-RegCM2, CCC, CCSR, CISRO-Mk2, GFDL, and HadCM3 general circulation model scenarios. High seasonal variability was also predicted within individual climatechange scenarios and large variability was indicated between scenarios within specific months. Overall, the climatechange scenarios reveal a large degree of uncertainty in current climatechange forecasts for the region. The results also indicate that the simu- lated UMRB hydrology is very sensitive to current forecasted futureclimate changes.
A major disadvantage of using GCMs is that they are considered to accurately represent global climate, but that they are often inaccurate when simulating regional climate (Evans and Schreider, 2002). GCMs are not well suited for answering the questions concerning regional- scale hydrological variability as they were not designed for this purpose due in part to their coarse spatial resolution, and general disagreement among GCMs with respect to precipitation changes at a regional scale (Loaiciga et al., 1996). With a coarse spatial resolution and the simplifications of the hydrologic cycle in those models, it is not possible to make reliable predictions of regional hydrologic changes directly from GCMs as a result of climate warming (Loaiciga et al., 1996). Downscaling or regionalization of GCM output is a common method of dealing with the regional inaccuracy. It should be noted that scenarios may play a key role in illustrating futureclimate changes in a region to help in strategic planning and water management decisions, but lack the confidence associated to allow a scenario to be referred to as a prediction or a forecast (Barrow et al., 2005).
The climatechangescenario provided by CanESM2 is available at a grid size of 2.8125°. The data available at this resolution is not suitable for hydrological analysis and thus highlights the necessity of downscaling so that it can be useful for basin scale analysis. GCMs in general are unable to resolve the sub-grid and regional climate scenarios and fail to take into account important regional features such as topography, vegetation, cloudiness which govern the local climate. Hence it is important to downscale the GCM data from global scale to local scale (Khadka et al. 2014). In the study, GCM outputs at global scale has been statistically downscaled to a point scale. Statistical down- scaling involves a technique in which a linear transfer fucntions between meso-scale atmospheric predictors var- iables of GCM (e.g. mean sea level pressure, geo-potential height, specific humidity etc.) and local climatic variables (temperature, precipitation etc.) are developed for ob- served period and these transfer functions are used to de- rive a point scale climate projections for future period. For this purpose, SDSM version 4.2 (Wilby and Dawson 2007) has been used for statistical downscaling. Several resera- chers have used SDSM for downscaling GCM data to a point scale (Khadka et al. 2014; Pervez and Henebry 2014; Babel et al. 2013; Mahmood and Babel 2013). CanESM2 provides GCM projections for three RCPs (RCP 2.6, RCP 4.5 & RCP 8.5) for the period of 2006 to 2100 along
The findings presented herein represent a deviation from the more broadly accepted viewpoint that forest disturbances will lead to reduced evapotranspiration and increased stream- flow. Previous findings tend to be based on studies carried out over short time periods (first 1–2 year responses, < 5 year studies), paired-basin analyses in watersheds disturbed by clear cuts, and where climate variability may have obscured results (Brown et al., 2005). Studies pointing towards in- creased streamflow also broadly found evapotranspiration from the canopy decreased, leading to an increase in runoff. However, work based on observations across scales and en- compassing multiple disturbances indicates that the regrowth potential for understory, such as shrubs used in this study, is high and that the regrowth is a major controlling factor for water availability and direction of change for evapotran- spiration and runoff (Caldwell et al., 2016; Biederman et al., 2014, 2015; Brown et al., 2014; Pribulick et al., 2016). Moreover, ecologists project that global forest covers are ex- pected to decline and be replaced with species and understory compositions that are more water intensive. It is therefore paramount to treat regrowth correctly within models to align these two currently disparate principles and account accu- rately for changing streamflow. This is a fundamental issue because ESMs rely upon the research principles developed on plot-scale and watershed-scale observational studies and modeling work.
a b s t r a c t
Study region: Upper Xingu RiverBasin, southeastern Amazonia.
Study focus: This study assessed the inﬂuence of land cover changes on evapo- transpiration and streamﬂow in small catchments in the Upper Xingu RiverBasin (Mato Grosso state, Brazil). Streamﬂow was measured in catchments with uniform land use for September 1, 2008 to August 31, 2010. We used mod- els to simulate evapotranspiration and streamﬂow for the four most common land cover types found in the Upper Xingu: tropical forest, cerrado (savanna), pasture, and soybean croplands. We used INLAND to perform single point sim- ulations considering tropical rainforest, cerrado and pasturelands, and AgroIBIS for croplands.
School of Water Resources and Environment, China University of Geosciences, Beijing, China Correspondence: Xu-Sheng Wang (email@example.com) and Sihai Liang (firstname.lastname@example.org) Received: 19 December 2017 – Revised: 6 February 2018 – Accepted: 11 February 2018 – Published: 5 June 2018 Abstract. Though extensive researches were conducted in the source region of the Yellow River (SRYR) to analyse climatechange influence on streamflow, however, few researches concentrate on streamflow of the sub- basin above the Huangheyan station in the SRYR (HSRYR) where a water retaining dam was built in the outlet in 1999. To improve the reservoir regulation strategies, this study analysed streamflowchange of the HSRYR in a mesoscale. A tank model (TM) was proposed and calibrated with monthly observation streamflow from 1991 to 1998. In the validation period, though there is a simulation deviation during the water storage and power generation period, simulated streamflow agrees favourably with observation data from 2008 to 2013. The model was further validated by two inside lakes area obtained from Landsat 5, 7, 8 datasets from 2000 to 2014, and significant correlations were found between the simulated lake outlet runoff and respective lake area. Then 21 Global Climate Models (GCM) ensembled data of three emission scenarios (SRA2, SRA1B and SRB1) were downscaled and used as input to the TM to simulate the runoff change of three benchmark periods 2011–2030 (2020s), 2046–2065 (2050s), 2080–2099 (2090s), respectively. Though temperature increase dramatically, these projected results similarly indicated that streamflow shows an increase trend in the long term. Runoff increase is mainly caused by increasing precipitation and decreasing evaporation. Water resources distribution is projected to change from summer-autumn dominant to autumn winter dominant. Annual lowest runoff will occur in May caused by earlier snow melting and increasing evaporation in March. According to the obtained results, winter runoff should be artificially stored by reservoir regulation in the future to prevent zero-flow occurrent in May. This research is helpful for water resources management and provides a better understand of streamflowchange caused by climatechange in the future.
Given that repeated DEM mapping may only be avail- able over periods of a decade or more, the approach applied here will require relatively long calibration periods – 15 yr in the case of Mica basin. The future projections (Fig. 11) demonstrate that changes in glacier area over decadal time scales can potentially have a significant influence on summer streamflow. Therefore, under conditions of rapid glacier re- treat, it may be necessary to represent the effects of glacier shrinkage during the calibration period, even in large catch- ments like Mica basin with modest glacier cover. Current long term planning data sets for the Columbia Region, used for example in hydroelectric operations planning, either do not account for glaciers (Hamlet et al., 2010) or assume static glaciers (Schnorbus et al., 2011; B¨urger et al., 2011). Our results suggest that, for climatechange impact assess- ments where glaciers are projected to recede substantially, the effects of glacier recession on streamflow have to be con- sidered even in basins with modest glacier cover (less than 10 %).
Assessment of direct human activities and climatechange on Weihe Riverstreamflow in this study showed that both factors had a significant effect. The streamflow and climatechange comparison results between the post-change period (1985-2010) and the pre-change period (1960-1984) for all four Weihe River hydrological stations are summarized in Table 5. From Table 5, annual streamflow in the post-change period obviously decreased, with predicted decreases of 51%, 49%, 36% and 39% at the Linjiacun, Xianyang, Lintong and Huaxian stations, respectively, as compared to pre-change period streamflow. The analysis showed that annual actual evaporation has slightly decreased during the post-change period, with percentage decreases ranging from 1% to 5% (Table 5). Annual precipitation during the post-change period was also estimated to have
April 1 SWE is a function of winter accumulation and ablation. SWE/P substantially decreases with each time period, indicating a hydrologic regime shift from a snow-rain-dominated to a rain-dominated basin. This is consistent with predictions in the Pacific Northwest [5,14,64–68]. Vynee et al.  predicted SWE to decrease more than 50% by the 2080s in the URB. A considerable change in basin area-weighted SWE has been observed to affect mid-elevation areas in the rain and snow transition zone . In post-fire conditions, there is a substantial decrease in SWE in the 2080s for both land cover conditions, a decrease of greater than 90%. This could be due to varying energy balances at the land and atmosphere interface, including radiative fluxes and changes in albedo, which can significantly influence the melting snow rate and the intensity of reflection by snow cover. Albedo was observed to be higher after a forest fire and lower after afforestation . Further analysis of montane snowpacks that store winter precipitation and provide water for the rest of the year is required for climate adaptation planning in dam water releases and flood control .
Stonefelt, Fontaine, and Hotchkiss (2000) and Boorman and Sefton (1997) both re- port results of +10 and -10 percent precipitation change scenarios for the Upper Wind RiverBasin and three United Kingdom watersheds ranging in size from 86 to 117 km 2 , respectively. Mean annual runoff impacts were predicted to range from about +16 to -15 percent in both studies, which were less than what was found in this study for the compa- rable Scenarios 4 and 5. The predicted decrease in water yield of over 50 percent for a 20 percent decline in precipitation (Scenario 3) was considerably higher than the 29 percent decrease in UMRB flows reported by Frederick (1993) for an analogue dust bowl cli- mate. His results were also influenced by the effects of higher temperature, which were incorporated into the analogue climatescenario. The effects of a 20 percent precipitation decrease (Scenario 3) simulated here (Table 5) were similar to seasonal flow impacts reported by Thomson et al. (2003) in response to El Niño conditions simulated for the UMRB, which ranged from -59 percent in summer to -33 percent in spring. Thomson et al. also report that a strong El Niño climate pattern was predicted to result in increased water yields ranging from 37 percent in summer to 62 percent in winter, which are similar to the percentage increases predicted in this study for Scenario 6 (Table 5). However, the largest flow increases were predicted to occur during the summer or fall in the present study, which essentially is the opposite of what Thomson et al. found. The Los Niños scenarios simulated by Thomson et al. also reflect the effects of temperature changes as well as precipitation fluctuations.
However, depending on the time-scale and performance measure of interest, different results arise regarding the model’s capability to adequately simulate streamflow under different precipitation forcing. When using the TRMMv7 precipitation product, the model is able to capture well the timing of the peaks and simulates well the recession limbs of the observed flows (Fig. 5.11). The simulation of rising limbs of the hydrograph is more problematic as often the increase in discharge is delayed compared to observations. Interestingly, to some extent, the results generated from forcing JULES with all different precipitation datasets seem to have similar behaviour, which indicates a model deficiency possibly related to the infiltration excess surface runoff generation mechanism. Clark and Gedney (2008) point out that delayed and lower flow peaks are the outcome of a runoff largely generated by drainage through the bot- tom of the soil column (i.e. subsurface runoff), as infiltration excess surface runoff occurs less frequently in large grid-scales. Besides, Dingman (2002) suggests that infiltration excess mostly occurs in arid and semi-arid regions where high intensity rainfall is combined with low surface conductivity, or in regions where human activity or frost has made the soil near-impermeable. The peak discharge is often underes- timated and in a smoothed hydrograph, of monthly or annual time scales, a clear low bias in the simulated streamflows is observed. This is especially noticeable in Fig. 5.13 that shows seasonally divided flows, where a strong underestimation of flows is observed during the wet summer season. Uncertainties in precipitation forcing datasets and to a lesser extent the runoff routing function might also be responsible for this mismatch. However, the simulations forced with IMD precipitation (Fig. 5.12) show problems in the rising limbs are almost minimised and especially in the monthly time scale, the achieved NSE is the highest of all. This indicates that precipitation is one of the most important factors that introduce uncertainties in the modelling. Unfortunately the time period of available gridded IMD precipitation data overlapping with observed flow data is short (1999–2004), therefore TRMMv7 was chosen as the forcing precipitation product from now on in this study.
Large-scale hydrological modelling does, however, involve many challenges [ 21 , 29 – 33 ]. The high spatial variability of input data such as land use, soil properties and topography across large catchments, the uncertainties in driving climate data, along with the difficulties of capturing micro-watershed scale hydrological processes and the ever-growing number of anthropogenic interventions incorporated at different periods in many river basins, directly affect the hydrological regime. This can make the accurate representation of river basins extremely difficult. For example, global runoff estimations can differ by as much as 70% between studies for individual continents [ 34 ]. Due to the highly influenced nature of many of India’s major river basins, including the presence of dams and water abstractions, the modelling of Indian water resources becomes extremely challenging [ 35 ]. The acquisition of reliable, relevant data poses a major challenge, where information on river flow and interventions is often not widely available. In light of this, a more holistic methodology and long-term assessment are needed for water resources management across many Indian catchments [ 36 ]. Large-scale model application in India ideally needs to incorporate anthropogenic basin interventions, such as water resource development projects, and account for population growth and demand from other water users [ 37 ], including industry and irrigated agriculture.
Predicted land cover map, LAI data and fixing the climatic conditions were used to drive the validated hydrological and sediment simulation models to investigate future land cover change impacts on streamflow and sediment load. Changes in streamflow in upstream and downstream are illustrated in Fig. 8, which indicate that land cover change affect stream- flow stronger in the downstream of DRB. As shown in Fig. 9, the effects of future land cover change would increase sed- iment load by 9.6 %, while the change rates in streamflow were within 5.7 %. Sediment load was found more sensitive to land cover change than streamflow. As for seasonally vari- ations of stream flow and sediment load caused by land cover changes, changes in streamflow and sediment load were more pronounced during the wet season.
- 68 -
flood with three scenarios in this case. However, for the scenario set 搬and the scenario set 斑 generated by
WG3 and WG-PCA3, the B11 (PCA) rule curves provide the best result for B11, B11(PCA), and historic(PCA) scenarios and the B21 rule curves represent the best results for B21 and B21(PCA) scenarios. These results demonstrate that the DE optimization procedure performs very well to determine the best adaptation strategy to the climatechange using existing storage in the basin. The total flood damages at each control point are decreased by the implementation of the optimal reservoir rule curves developed for various climatechange scenarios. In addition, the WG-PCA3 provides more wet weather conditions than the original WG model.
Abstract In this study, a hydrological modelling framework was introduced to assess the climatechange impacts on futureriver flow in the West Riverbasin, China, especially on streamflow variability and extremes. The modelling framework includes a delta-change method with the quantile-mapping technique to construct futureclimate forcings on the basis of observed meteorological data and the downscaled climate model outputs. This method is able to retain the signals of extreme weather events, as projected by climate models, in the constructed future forcing scenarios. Fed with the historical and future forcing data, a large- scale hydrologic model (the Variable Infiltration Capacity model, VIC) was executed for streamflow simulations and projections at daily time scales. A bootstrapping resample approach was used as an indirect alternative to test the equality of means, standard deviations and the coefficients of variation for the baseline and futurestreamflow time series, and to assess the future changes in flood return levels. The West Riverbasin case study confirms that the introduced modelling framework is an efficient effective tool to quantify streamflow variability and extremes in response to futureclimatechange.
In drylands, convective rainstorms typically control runoff, streamflow, water supply and flood risk to human populations, and ecological water availability at multiple spatial scales. Since drainage basin water balance is sensitive to climate, it is important to improve characterization of convective rainstorms in a manner that enables statistical assessment of rainfall at high spatial and temporal resolution, and the prediction of plausible manifestations of climatechange. Here we present a simple rainstorm generator, STORM, for convective storm simulation. It was created using data from a rain gauge network in one dryland drainage basin, but is applicable anywhere. We employ STORM to assess watershed rainfall under climatechange simulations that reflect differences in wetness/ storminess, and thus provide insight into observed or projected regional hydrologic trends. Our analysis documents historical, regional climatechange manifesting as a multidecadal decline in rainfall intensity, which we suggest has negatively impacted ephemeral runoff in the Lower ColoradoRiverbasin, but has not contributed substantially to regional negative streamflow trends.