Hydrological simulation of the Betwa River basin (India) using the SWAT model

A seasonal water budget analysis was carried out to quantify various components of the hydrological cycle using the Soil and Water Assessment Tool (SWAT) model for the Betwa River basin (43?500 km²) in central India. The model results were satisfactory in calibration and validation. The seasonal water budget analysis showed that about 90% of annual rainfall and 97% of annual runoff occurred in the monsoon season. A seasonal linear trend analysis was carried out to detect trends in the water balance components of the basin for the period 1973–2001. In the monsoon season, an increasing trend in rainfall and a decreasing trend in ET were observed; this resulted in an increasing trend in groundwater storage and surface runoff. The winter season followed almost the same pattern. A decreasing trend was observed in summer season rainfall. The study evokes the need for conservation structures in the study area to reduce monsoon runoff and conserve it for basin requirements in water-scarce seasons.

EDITOR Z.W. Kundzewicz

ASSOCIATE EDITOR F. Hattermann     

Distributed snow accumulation and ablation modeling in the American River basin

A spatially distributed snow model procedure for estimating snow melt, snow water equivalent and snow cover area is formulated and tested with data from the American River basin in California’s Sierra Nevada. An adaptation of the operational National Weather Service snow accumulation and ablation model is used for each model grid cell forced by spatially distributed precipitation and temperature data. The model was implemented with 6-hourly time steps on 1 km² grid cells for the snow season of 1999–2003. Temperature is spatially interpolated using the prevailing lapse rate and digital terrain elevation data. Precipitation is spatially interpolated using regional climatological analyses obtained from PRISM. Parameters that control snow melt are distributed using ground surface aspect. The model simulations are compared with data from 12 snow-sensors located in the basin and the daily 500-m snow cover extent product from the MODIS/Terra satellite mission. The results show that the distribution of snow pack over the area is generally captured. The snow pack quantity compared to snow gauges is well estimated in high elevations with increasing uncertainty in the snow pack at lower elevations. Sensitivity and uncertainty analyses indicate that the significant input uncertainty for precipitation and temperature is primarily responsible for model errors in lower elevations and near the snow line. The model is suitable for producing spatially resolved realistic snow pack simulations when forced with operationally available observed or predicted data.     

Estimating the distribution of snow water equivalent and snow extent beneath cloud cover in the Salt–Verde River basin,Arizona

The temporal and spatial continuity of spatially distributed estimates of snow‐covered area (SCA) are limited by the availability of cloud‐free satellite imagery; this also affects spatial estimates of snow water equivalent (SWE), as SCA can be used to define the extent of snow telemetry (SNOTEL) point SWE interpolation. In order to extend the continuity of these estimates in time and space to areas beneath the cloud cover, gridded temperature data were used to define the spatial domain of SWE interpolation in the Salt–Verde watershed of Arizona. Gridded positive accumulated degree‐days (ADD) and binary SCA (derived from the Advanced Very High Resolution Radiometer (AVHRR)) were used to define a threshold ADD to define the area of snow cover. The optimized threshold ADD increased during snow accumulation periods, reaching a peak at maximum snow extent. The threshold then decreased dramatically during the first time period after peak snow extent owing to the low amount of energy required to melt the thin snow cover at lower elevations. The area having snow cover at this later time was then used to define the area for which SWE interpolation was done. The area simulated to have snow was compared with observed SCA from AVHRR to assess the simulated snow map accuracy. During periods without precipitation, the average commission and omission errors of the optimal technique were 7% and 11% respectively, with a map accuracy of 82%. Average map accuracy decreased to 75% during storm periods, with commission and omission errors equal to 11% and 12% respectively. The analysis shows that temperature data can be used to help estimate the snow extent beneath clouds and therefore improve the spatial and temporal continuity of SCA and SWE products. Copyright © 2004 John Wiley & Sons, Ltd.     

Hydrological analysis of the Upper Tiber River Basin,Central Italy: a watershed modelling approach

The quantification of the various components of hydrological processes in a watershed remains a challenging topic as the hydrological system is altered by internal and external drivers. Watershed models have become essential tools to understand the behaviour of a catchment under dynamic processes. In this study, a physically based watershed model called Soil Water Assessment Tool was used to understand the hydrologic behaviour of the Upper Tiber River Basin, Central Italy. The model was successfully calibrated and validated using observed weather and flow data for the period of 1963–1970 and 1971–1978, respectively. Eighteen parameters were evaluated, and the model showed high relative sensitivity to groundwater flow parameters than the surface flow parameters. An analysis of annual hydrological water balance was performed for the entire upper Tiber watershed and selected subbasins. The overall behaviour of the watershed was represented by three categories of parameters governing surface flow, subsurface flow and whole basin response. The base flow contribution has shown that 60% of the streamflow is from shallow aquifer in the subbasins. The model evaluation statistics that evaluate the agreement between the simulated and the observed streamflow at the outlet of a watershed and other three different subbasins has shown a coefficient of determination (R²) from 0.68 to 0.81 and a Nash–Sutcliffe efficiency (E_NS) between 0.51 and 0.8 for the validation period. The components of the hydrologic cycle showed variation for dry and wet periods within the watershed for the same parameter sets. On the basis of the calibrated parameters, the model can be used for the prediction of the impact of climate and land use changes and water resources planning and management. Copyright © 2012 John Wiley & Sons, Ltd.     

Long‐term investigations of the snow cover in a subalpine semi‐forested catchment

To improve spring runoff forecasts from subalpine catchments, detailed spatial simulations of the snow cover in this landscape is obligatory. For more than 30 years, the Swiss Federal Research Institute WSL has been conducting extensive snow cover observations in the subalpine watershed Alptal (central Switzerland). This paper summarizes the conclusions from past snow studies in the Alptal valley and presents an analysis of 14 snow courses located at different exposures and altitudes, partly in open areas and partly in forest. The long‐term performance of a physically based numerical snow–vegetation–atmosphere model (COUP) was tested with these snow‐course measurements. One single parameter set with meteorological input variables corrected to the prevailing local conditions resulted in a convincing snow water equivalent (SWE) simulation at most sites and for various winters with a wide range of snow conditions. The snow interception approach used in this study was able to explain the forest effect on the SWE as observed on paired snow courses. Finally, we demonstrated for a meadow and a forest site that a successful simulation of the snowpack yields appropriate melt rates. Copyright © 2006 John Wiley & Sons, Ltd.     

Seasonal snow cover decreases young water fractions in high Alpine catchments

Estimation of young water fractions (F_yw), defined as the fraction of water in a stream younger than approximately 2–3 months, provides key information for water resource management in catchments where runoff is dominated by snowmelt. Knowing the average dependence of summer flow on winter precipitation is an essential context for comparing regional drought severity and provides the hydrological template for downstream water users and ecosystems. However, F_yw estimation based on seasonal signals of stable isotopes of oxygen and hydrogen has not yet explicitly addressed how to parsimoniously include the seasonal shift of water input from snow. Using experimental data from three high-elevation, Alpine catchments (one dominated by glacier and two by snow), we propose a framework to explicitly include the delays induced by snow storage into estimates of F_yw. Scrutinizing the key methodological choices when estimating F_yw from isotope data, we find that the methods used to construct precipitation input signals from sparse isotope samples can significantly impact F_yw. Given this sensitivity, our revised procedure estimates a distribution of F_yw values that incorporates a wide range of possible methodological choices and their uncertainties; it furthermore compares the commonly used amplitude ratio approach to a direct convolution approach, which circumvents the assumption that the isotopic signals have a sine curve shape, an assumption that is generally violated in snow-dominated environments. Our new estimates confirm that high-elevation Alpine catchments have low F_yw values, spanning from 8 to 11%. Such low values have previously been interpreted as the impact of seasonal snow storage alone, but our comparison of different F_yw estimation methods suggests that these low F_yw values result from a combination of both snow cover effects and longer storage in the subsurface. In contrast, in the highest elevation, glacier dominated catchment, F_yw is 3–4 times greater compared to the other two catchments, due to the lower storage and faster drainage processes. A future challenge, capturing spatio-temporal snowmelt isotope signals during winter baseflow and the snowmelt period, remains to improve constraints on the F_yw estimation technique.     

Snow water equivalent estimation in the Mallero basin using snow gauge data and MODIS images and fieldwork validation

Snow accumulation and ablation rule the temporal dynamics of water availability in mountain areas and cold regions. In these environments, the evaluation of the snow water amount is a key issue. The spatial distribution of snow water equivalent (SWE) over a mountain basin at the end of the snow accumulation season is estimated using a minimal statistical model (SWE‐SEM). This uses systematic observations such as ground measurements collected at snow gauges and snow‐covered area (SCA) data retrieved by remote sensors, here MODIS. Firstly, SWE‐SEM calculates local SWE estimates at snow gauges, then the spatial distribution of SWE over a certain area using an interpolation method; linear regressions of the first two order moments of SWE with altitude. The interpolation has been made by both confining and unconfining the spatial domain by SCA. SWE‐SEM is applied to the Mallero basin (northern Italy) for calculating the snow water equivalent at the end of the winter season for 6 years (2001–2007). For 2007, SWE‐SEM estimates are validated through fieldwork measurements collected during an ‘ad hoc’ campaign on March 31, 2007. Snow‐surveyed measurements are used to check SCA, snow density and SWE estimates. Copyright © 2009 John Wiley & Sons, Ltd.     

Subgrid variability of snow water equivalent at operational snow stations in the western USA

The spatial distribution of snow water equivalent (SWE) is a key variable in many regional‐scale land surface models. Currently, the assimilation of point‐scale snow sensor data into these models is commonly performed without consideration of the spatial representativeness of the point data with respect to the model grid‐scale SWE. To improve the understanding of the relationship between point‐scale snow measurements and surrounding areas, we characterized the spatial distribution of snow depth and SWE within 1‐, 4‐ and 16‐km² grids surrounding 15 snow stations (snowpack telemetry and California snow sensors) in California, Colorado, Wyoming, Idaho and Oregon during the 2008 and 2009 snow seasons. More than 30 000 field observations of snowpack properties were used with binary regression tree models to relate SWE at the sensor site to the surrounding area SWE to evaluate the sensor representativeness of larger‐scale conditions. Unlike previous research, we did not find consistent high biases in snow sensor depth values as biases over all sites ranged from 74% overestimates to 77% underestimates. Of the 53 assessments, 27 surveys indicated snow station biases of less than 10% of the surrounding mean observed snow depth. Depth biases were largely dictated by the physiographic relationship between the snow sensor locations and the mean characteristics of the surrounding grid, in particular, elevation, solar radiation index and vegetation density. These scaling relationships may improve snow sensor data assimilation; an example application is illustrated for the National Operational Hydrologic Remote Sensing Center National Snow Analysis SWE product. The snow sensor bias information indicated that the assimilation of point data into the National Operational Hydrologic Remote Sensing Center model was often unnecessary and reduced model accuracy. Copyright © 2012 John Wiley & Sons, Ltd.     

Assessment of surface water resources and evapotranspiration in the Haihe River basin of China using SWAT model

Quantitative assessment of surface water resources (SWRs) and evapotranspiration (ET) is essential and significant for reasonably planning and managing water resources in the Haihe River basin which is facing severe water shortage. In this study, a distributed hydrological model of the Haihe River basin was constructed using the Soil and Water Assessment Tool, well considering the reservoirs and agricultural management practices for reasonable simulation. The crop parameters were independently calibrated with the observed crop data at six experimental stations. Then, sensitivity ranks of hydrological parameters were analysed, which suggested the important parameters used for calibration. The model was successfully calibrated using the monthly observed data of discharge in around 1970–1991 and actual ET (ET_a) in 2002–2004 for the mountainous area and Haihe plain, respectively. Meanwhile, good agreements between the simulated and statistical crop yields in 1985–2005 further verified the model's appropriateness. Finally, the calibrated model was used to assess SWRs and ET_a in time and space during 1961–2005. Results showed that the average annual natural SWRs and the ET_a were about 17.5 billion cubic metre and 542 mm, respectively, both with a slight downward trend. The spatial distributions of both SWRs and ET_a were significantly impacted by variations of precipitation and land use. Moreover, the reservoir in operation was the main factor for the noticeable decline of actual SWRs. In the Haihe plain, the ET_a with irrigation was increased by 46% compared with that under rainfed conditions. In addition, this study identified the regions with potential to improve the irrigation effects on water use. Copyright © 2012 John Wiley & Sons, Ltd.     

Melted snow volume control in the snowmelt runoff model using a snow water equivalent statistically based model

Snowmelt is an important component of the river discharge in mountain environments. In the past 40 years, the snowmelt dynamics has been mostly evaluated using degree‐day‐based models like the snowmelt runoff model (SRM). This model has no control on the volume of the melting snow, even if SRM includes as data input the snow‐covered area. This lack explains why the application of SRM may lead to inaccurate snowmelt volume estimations, even if the discharge volumes are accurately reproduced. Here we introduce in SRM the control on the melted snow volume and consider it in the determination of SRM parameters. The total snow volume, accumulated at the end of winter season, is evaluated by a snow water equivalent statistically based model, SWE‐SEM, and used as an estimate of the melting snow during the summer season. The benefit derived from the introduction of the control on the melting snow volume was investigated in the Mallero basin (northern Italy) for the 2003 and 2004 snow melting seasons. The analysis compares the model's results adopting different parameter sets, both considering and ignoring the control on the melting snow volume. Copyright © 2011 John Wiley & Sons, Ltd.     

The role of meteorological forcing and snow model complexity in winter glacier mass balance estimation,Columbia River basin,Canada

Glaciers are commonly located in mountainous terrain subject to highly variable meteorological conditions. High resolution meteorological (HRM) data simulated by atmospheric models can complement meteorological station observations in order to assess changes in glacier energy fluxes and mass balance. We examine the performance of two snow models, SnowModel and Alpine3D, forced by different meteorological data for winter mass balance simulations at four glaciers in the Canadian portion of the Columbia Basin. The Weather Research and Forecasting model (WRF) with resolution of 1 km and the North American Land Data Assimilation System with ~12 km resolution, provide HRM data for the two snow models. Evaluation is based on the ability of the snow models to simulate snow depth at both point locations (automated snow weather stations) and over the entire glacier surface (airborne LiDAR [Light Detection and Ranging] surveys) during the 2015/2016 winter accumulation. When forced with HRM data, both models can reproduce snow depth to within ±15% of observed values. Both models underestimate winter mass balance when forced by HRM data. When driven with WRF data, SnowModel underestimates winter mass balance integrated over the glacier area by 1 and 10%, whilst Alpine3D underestimates winter mass balance by 12 and 22% compared with LiDAR and stake measurements, respectively. The overall results show that SnowModel forced by WRF simulated winter mass balance the best.     

Attribution of changes in the water balance of a tropical catchment to land use change using the SWAT model

Changes in the water balance of the Samin catchment (277.9 km²) on Java, Indonesia, can be attributed to land use change using the Soil Water Assessment Tool model. A baseline‐altered method was used in which the simulation period 1990–2013 was divided into 4 equal periods to represent baseline conditions (1990–1995) and altered land use conditions (1996–2001, 2002–2007, and 2008–2013). Land use maps for 1994, 2000, 2006, and 2013 were acquired from satellite images. A Soil Water Assessment Tool model was calibrated for the baseline period and applied to the altered periods with and without land use change. Incorporating land use change resulted in a Nash–Sutcliffe efficiency of 0.7 compared to 0.6 when land use change is ignored. In addition, the model performance for simulations without land use change gradually decreased with time. Land use change appeared to be the important driver for changes in the water balance. The main land use changes during 1994–2013 are a decrease in forest area from 48.7% to 16.9%, an increase in agriculture area from 39.2% to 45.4%, and an increase in settlement area from 9.8% to 34.3%. For the catchment, this resulted in an increase of the runoff coefficient from 35.7% to 44.6% and a decrease in the ratio of evapotranspiration to rainfall from 60% to 54.8%. More pronounced changes can be observed for the ratio of surface runoff to stream flow (increase from 26.6% to 37.5%) and the ratio of base flow to stream flow (decrease from 40% to 31.1%), whereas changes in the ratio of lateral flow to stream flow were minor (decrease from 33.4% to 31.4%). At sub‐catchment level, the effect of land use changes on the water balance varied in different sub‐catchments depending on the scale of changes in forest and settlement area.     

Evaluation of snow models in terrestrial biosphere models using ground observation and satellite data: impact on terrestrial ecosystem processes

Snow is important for water management, and an important component of the terrestrial biosphere and climate system. In this study, the snow models included in the Biome‐BGC and Terrestrial Observation and Prediction System (TOPS) terrestrial biosphere models are compared against ground and satellite observations over the Columbia River Basin in the US and Canada and the impacts of differences in snow models on simulated terrestrial ecosystem processes are analysed. First, a point‐based comparison of ground observations against model and satellite estimates of snow dynamics are conducted. Next, model and satellite snow estimates for the entire Columbia River Basin are compared. Then, using two different TOPS simulations, the default TOPS model (TOPS with TOPS snow model) and the TOPS model with the Biome‐BGC snow model, the impacts of snow model selection on runoff and gross primary production (GPP) are investigated. TOPS snow model predictions were consistent with ground and satellite estimates of seasonal and interannual variations in snow cover, snow water equivalent, and snow season length; however, in the Biome‐BGC snow model, the snow pack melted too early, leading to extensive underpredictions of snow season length and snow covered area. These biases led to earlier simulated peak runoff and reductions in summer GPP, underscoring the need for accurate snow models within terrestrial ecosystem models. Copyright © 2007 John Wiley & Sons, Ltd.     

Improving the SWAT forest module for enhancing water resource projections: A case study in the St. Croix River basin

Insufficiently calibrated forest parameters of the Soil & Water Assessment Tool (SWAT) may introduce uncertainties to water resource projections in forested watersheds. In this study, we improved SWAT forest parameterization and phosphorus cycling representations to better simulate forest ecosystems in the St. Croix River basin, and we further examined how those improvements affected model projections of streamflow, sediment, and nitrogen export under future climate conditions. Simulations with improved forest parameters substantially reduced model estimates of water, sediment, and nitrogen fluxes relative to those based on default parameters. Differences between improved and default projections can be attributed to the enhanced representation of forest water consumption, nutrient uptake, and protection of soil from erosion. Better representation of forest ecosystems in SWAT contributes to constraining uncertainties in water resource projections. Results of this study highlight the importance of improving SWAT forest ecosystem representations in projecting delivery of water, sediment, and nutrients from land to rivers in response to climate change, particularly for watersheds with large areas of forests. Improved forest parameters and the phosphorus weathering algorithms developed in this study are expected to help enhance future applications of SWAT to investigate hydrological and biogeochemical consequences of climate change.     

Evaluation of snow water equivalent datasets over the Saint‐Maurice river basin region of southern Québec

A 10‐km gridded snow water equivalent (SWE) dataset is developed over the Saint‐Maurice River basin region in southern Québec from kriging of observed snow survey data for evaluation of SWE products. The gridded SWE dataset covers 1980–2014 and is based on manual gravimetric snow surveys carried out on February 1, March 1, March 15, April 1, and April 15 of each snow season, which captures the annual maximum SWE (SWEM) with a mean interpolation error of ±19%. The dataset is used to evaluate SWEM from a range of sources including satellite retrievals, reanalyses, Canadian regional climate models, and the Canadian Meteorological Centre operational snow depth analysis. We also evaluate a number of solid precipitation datasets to determine their contribution to systematic errors in estimated SWEM. None of the evaluated datasets is able to provide estimates of SWEM that are within operational requirements of ±15% error, and insufficient solid precipitation is determined to be one of the main reasons. The Climate System Forecast Reanalysis is the only dataset where snowfall is sufficiently large to generate SWEM values comparable to observations. Inconsistencies in precipitation are also found to have a strong impact on year‐to‐year variability in SWEM dataset performance and spread. Version 3.6.1 of the Canadian Land Surface Scheme land surface scheme driven with ERA‐Interim output downscaled by Version 5.0.1 of the Canadian Regional Climate Model was the best physically based model at explaining the observed spatial and temporal variability in SWEM (root‐mean‐square error [RMSE] = 33%) and has potential for lower error with adjusted precipitation. Operational snow products relying on the real‐time snow depth observing network performed poorly due to a lack of real‐time data and the strong local scale variability of point snow depth observations. The results underscore the need for more effort to be invested in improving solid precipitation estimates for use in snow hydrology applications.     

Determination of snow water equivalent over the eastern part of Turkey using passive microwave data

Information on regional snow water equivalent (SWE) is required for the management of water generated from snowmelt. Modeling of SWE in the mountainous regions of eastern Turkey, one of the major headwaters of Euphrates–Tigris basin, has significant importance in forecasting snowmelt discharge, especially for optimum water usage. An assimilation process to produce daily SWE maps is developed based on Helsinki University of Technology (HUT) model and AMSR‐E passive microwave data. The characteristics of the HUT emission model are analyzed in depth and discussed with respect to the extinction coefficient function. A new extinction coefficient function for the HUT model is proposed to suit models for snow over mountainous areas. Performance of the modified model is checked against the original, other modified cases and ground truth data covering the 2003–2007 winter periods. A new approach to calculate grain size and density is integrated inside the developed data assimilation process. An extensive validation was successfully performed by means of snow data measured at ground stations during the 2008–2010 winter periods. The root mean square error of the data set for snow depth and SWE between January and March of the 2008–2010 periods compared with the respective AMSR‐E footprints indicated that errors for estimated snow depth and predicted SWE values were 16.92 cm and 40.91 mm, respectively, for the 3‐year period. Validation results were less satisfactory for SWE less than 75.0 mm and greater than 150.0 mm. An underestimation for SWE greater than 150 mm could not be resolved owing to the microwave signal saturation that is observed for dense snowpack. Copyright © 2012 John Wiley & Sons, Ltd.     

Modelling the impact of agroforestry on hydrology of Mara River Basin in East Africa

Land‐use change is one of the main drivers of watershed hydrology change. The effect of forestry related land‐use changes (e.g. afforestation, deforestation, agroforestry) on water fluxes depends on climate, watershed characteristics and spatial scale. The Soil and Water Assessment Tool (SWAT) model was calibrated, validated and used to simulate the impact of agroforestry on the water balance in the Mara River Basin (MRB) in East Africa. Model performance was assessed by Nash–Sutcliffe Efficiency (NSE) and Kling–Gupta Efficiency (KGE). The NSE (and KGE) values for calibration and validation were: 0.77 (0.88) and 0.74 (0.85) for the Nyangores sub‐watershed, and 0.78 (0.89) and 0.79 (0.63) for the entire MRB. It was found that agroforestry in the watershed would generally reduce surface runoff, mainly because of enhanced infiltration. However, it would also increase evapotranspiration and consequently reduce baseflow and overall water yield, which was attributed to increased water use by trees. Spatial scale was found to have a significant effect on water balance; the impact of agroforestry was higher at the smaller headwater catchment (Nyangores) than for the larger watershed (entire MRB). However, the rate of change in water yield with an increase in area under agroforestry was different for the two and could be attributed to the spatial variability of climate within the MRB. Our results suggest that direct extrapolation of the findings from a small sub‐catchment to a larger watershed may not always be accurate. These findings could guide watershed managers on the level of trade‐offs that might occur between reduced water yields and other benefits (e.g. soil erosion control, improved soil productivity) offered by agroforestry. Copyright © 2016 John Wiley & Sons, Ltd.     

Preliminary study on assessment of nutrient transport in the Taihu Basin based on SWAT modeling

With the Taihu Basin as a study area, using the spatially distributed and mechanism-based SWAT model, preliminary simulations of nutrient transport in the Taihu Basin during the period of 1995-2002 has been carried out. The topography, soil, meteorology and land use with industrial point pollution discharge, the loss of agricultural fertilizers, urban sewerage, and livestock drainages were all considered in the boundary conditions of the simulations. The model was calibrated and validated against water quality monitoring data from 2001 to 2002. The results show that the annual total productions of nitrogen (TN) and phosphorus (TP) into Lake Taihu are 40000t and 2000t respectively. Nutrient from the Huxi Region is a major resource for Lake Taihu. The non-point source (surface source) pollution is the main form of catchment sources of nutrients into Lake Taihu, occupied TN 53% and TP 56% respectively. TN and TP nutrients from industrial point pollution discharge are 30% and 16%, and sewerage in both forms of point source and non-point source are TN 31 % and TP 47%. Both the loss of agricultural fertilizers and livestock drainages from the catchment should be paid more attention as an important nutrient source. The results also show that SWAT is an effective model for the simulation of temporally and spatially nutrient changes and for the assessment of the trends in a catchment scale.     

Performance of temperature and radiation index models for point-scale snow water equivalent (SWE) simulations in the Moroccan High Atlas Mountains

ABSTRACT

Four models of increasing complexity were tested and compared to simulate snow water equivalent at the local scale in the Moroccan High Atlas range. A classical temperature index model (TI) and three enhanced temperature index models that respectively include the potential clear-sky direct radiation (HTI), the incoming solar radiation (ETI-A) and net solar radiation (ETI-B), were subjected to annual and multi-annual calibration and cross-validated over the period 2003–2010. When calibrated yearly, the ETI models could be better transferred to other years, whereas all models, including the simple TI model, showed good transferability when calibrated over a longer period that includes inter-annual climate variability. No strong and recurrent relationships emerged between yearly calibrated model parameters and annual climate conditions. However, strong parameter compensation was observed for the enhanced models, which can be explained partly by the collinearity of air temperature and solar radiation causing equifinality of model parameters.