Hydrological Tools Research Articles

Watershed Delineation in Flat Terrain of Thar Desert Region in North West India – A Semi Automated Approach Using DEM

Thar desert spreads in western part of Rajasthan, northern part of Gujarat, and some parts of Punjab and Haryana. The terrain is dominated by slightly sloping plains, broken by some dunes and low barren hills. The area is characterized by low average annual rainfall which is erratic in distribution and intensity. Drought will remain a major hindrance for agricultural production in Thar desert. Due to water stress condition, many watershed based development activities has been adopted by government and non-government organizations for the growth and sustainable development of this region. The need of this hour is preparation of a national level watershed atlas of 1:50,000 scale because majority of thematic maps are being produced presently on same or 1:10,000 scale. The manual delineation of watershed boundary in flat terrain based on topographic map will be time consuming and less accurate in the absence of prominent contour lines. Automated approach for watershed delineation using Digital Elevation Model (DEM) along a suitable algorithm has the advantage because the output is not only less time consuming but also independent from human decisions. Hence, a case study has been carried out in Churu sub-basin part of Indus basin which is located in Thar desert region. Depression less DEM with different spatial resolutions was used as input in hydrology tool of ArcGIS spatial analyst function for characterization of watersheds. The Churu sub-basin has been divided into various numbers of watersheds with an average size of 600 km2. These watershed boundaries have been validated with respect to high resolution satellite imageries (IRS P6 LISS IV), Survey of India toposheets, ancillary data and limited field checks.

Extraction of a representative instantaneous unit hydrograph

This paper deals with the problem of the presence of noise in the calculated instantaneous unit hydrograph (IUH). Daubechies wavelets are used to decompose and smooth the derived IUH by way of fast Fourier transform. The results reveal the effectiveness of wavelet analysis in obtaining a fine separation of the noise frequencies from the main frequencies relating to IUHs. Choosing a representative IUH through the smoothed IUHs by a cross-validation technique, it can be applied to predict other flood events or as a design tool in hydrology.

Exploiting remote sensing land surface temperature in distributed hydrological modelling: the example of the Continuum model

Abstract. Full process description and distributed hydrological models are very useful tools in hydrology as they can be applied in different contexts and for a wide range of aims such as flood and drought forecasting, water management, and prediction of impact on the hydrologic cycle due to natural and human-induced changes. Since they must mimic a variety of physical processes, they can be very complex and with a high degree of parameterization. This complexity can be increased by necessity of augmenting the number of observable state variables in order to improve model validation or to allow data assimilation. In this work a model, aiming at balancing the need to reproduce the physical processes with the practical goal of avoiding over-parameterization, is presented. The model is designed to be implemented in different contexts with a special focus on data-scarce environments, e.g. with no streamflow data. All the main hydrological phenomena are modelled in a distributed way. Mass and energy balance are solved explicitly. Land surface temperature (LST), which is particularly suited to being extensively observed and assimilated, is an explicit state variable. A performance evaluation, based on both traditional and satellite derived data, is presented with a specific reference to the application in an Italian catchment. The model has been firstly calibrated and validated following a standard approach based on streamflow data. The capability of the model in reproducing both the streamflow measurements and the land surface temperature from satellites has been investigated. The model has been then calibrated using satellite data and geomorphologic characteristics of the basin in order to test its application on a basin where standard hydrologic observations (e.g. streamflow data) are not available. The results have been compared with those obtained by the standard calibration strategy based on streamflow data.

Water Resources Sustainability Indicator: Application of the Watershed Characteristics Approach

The quantification of the renewable flux (i.e. sustainable limit) of the hydrologic system is the prerequisite for transitioning from unsustainable to sustainable water resources management. The application of the Watershed Characteristics Approach to estimate the renewable flux of the hydrologic system was demonstrated using Minnesota’s (USA) Twin Cities Metropolitan Area (TCMA). The methodology quantified the relationships between landscape properties and water balance characteristics, resulting in the development of functioning hierarchical hydrogeological units with corresponding recharge rates. This renewable flux is a key quantitative characteristic for the assessment of a sustainability indicator. The key indicator of sustainable water use is the ratio of the renewable capacity of the hydrologic system to the water use by humans and the environment. By incorporating water use estimates for the TCMA relative to the calculated recharge rates, sustainability indicators for groundwater and total flux were calculated for the metropolitan area. As far back as the 1890s, declines in TCMA groundwater levels have been observed, which correspond to the unsustainable groundwater extraction estimates identified in the results of this study. The non-stationary characteristics of urban watersheds influenced by ongoing land use/land cover changes as illustrated in this paper, emphasizes the need for conservative hydrologic planning to achieve sustainable water management. This approach can also be applied to other metropolitan areas as a hydrologic tool for decision-makers to design sustainable water policy and prevent the over-extraction of the water flowing through the hydrologic system.

Mapping the elevation change of Lambert Glacier in East Antarctica using ICESat GLAS

We initially derived elevation changes of Geoscience Laser Altimeter System level-2 altimetry data of Lambert Glacier overlapping footprints during each mission from 2003 to 2008. Then, surface elevation changes during every two adjacent missions were interpolated using inverse distance weighted, natural neighbor, triangulated irregular network with linear method, radial basis functions and ANUDEM in ArcGIS. The best results were obtained by ANUDEM, so these data were clipped to conform to the study area boundary defined by hydrology tools. Finally, elevation changes over 10 periods were mapped. In these maps, we chose the Antarctic digital elevation model as background and used a translucent layer to mask the area outside Lambert Glacier, and then displayed elevation changes using gradient colors. Results indicate that elevation changes of the entire Lambert Glacier are not evident, particularly in the upstream area. There are a few elevation changes in some downstream areas. Elevation of the grounding zone in the southernmost Amery Ice Shelf decreased more than 2 m in 2004–2005, 2006–2007, and during 2008.

A comparison of methods to avoid overfitting in neural networks training in the case of catchment runoff modelling

Summary Artificial neural networks (ANNs) becomes very popular tool in hydrology, especially in rainfall–runoff modelling. However, a number of issues should be addressed to apply this technique to a particular problem in an efficient way, including selection of network type, its architecture, proper optimization algorithm and a method to deal with overfitting of the data. The present paper addresses the last, rarely considered issue, namely comparison of methods to prevent multi-layer perceptron neural networks from overfitting of the training data in the case of daily catchment runoff modelling. Among a number of methods to avoid overfitting the early stopping, the noise injection and the weight decay have been known for about two decades, however only the first one is frequently applied in practice. Recently a new methodology called optimized approximation algorithm has been proposed in the literature. Overfitting of the training data leads to deterioration of generalization properties of the model and results in its untrustworthy performance when applied to novel measurements. Hence the purpose of the methods to avoid overfitting is somehow contradictory to the goal of optimization algorithms, which aims at finding the best possible solution in parameter space according to pre-defined objective function and available data. Moreover, different optimization algorithms may perform better for simpler or larger ANN architectures. This suggest the importance of proper coupling of different optimization algorithms, ANN architectures and methods to avoid overfitting of real-world data – an issue that is also studied in details in the present paper. The study is performed for Annapolis River catchment, characterized by significant seasonal changes in runoff, rapid floods during winter and spring, moderately dry summers, severe winters with snowfall, snow melting, frequent freeze and thaw, and presence of river ice. The present paper shows that the elaborated noise injection method may prevent overfitting slightly better than the most popular early stopping approach. However, the implementation of noise injection to real-world problems is difficult and the final model performance depends significantly on a number of very technical details, what somehow limits its practical applicability. It is shown that optimized approximation algorithm does not improve the results obtained by older methods, possibly due to over-simplified criterion of stopping the algorithm. Extensive calculations reveal that Evolutionary Computation-based algorithm performs better for simpler ANN architectures, whereas classical gradient-based Levenberg–Marquardt algorithm is able to benefit from additional input variables, representing precipitation and snow cover from one more previous day, and from more complicated ANN architectures. This confirms that the curse of dimensionality has severe impact on the performance of Evolutionary Computing methods.

Hydrologic information server for benchmark precipitation dataset

This paper will present the methodology and overall system development by which a benchmark dataset of precipitation information has been made available. Rainfall is the primary driver of the hydrologic cycle. High quality precipitation data is vital for hydrologic models, hydrometeorologic studies and climate analysis,and hydrologic time series observations are important to many water resources applications. Over the past two decades, with the advent of NEXRAD radar, science to measure and record rainfall has improved dramatically. However, much existing data has not been readily available for public access or transferable among the agricultural, engineering and scientific communities. This project takes advantage of the existing CUAHSI Hydrologic Information System ODM model and tools to bridge the gap between data storage and data access, providing an accepted standard interface for internet access to the largest time-series dataset of NEXRAD precipitation data ever assembled. This research effort has produced an operational data system to ingest, transform, load and then serve one of most important hydrologic variable sets.

Geomorphology-based Time-Lagged Recurrent Neural Networks for runoff forecasting

Artificial Neural Networks have been widely used to develop effective runoff-forecasting models. An overwhelming majority of networks are static in nature and also developed without incorporating geomorphologic information of the watershed. The objective of this study is to develop an efficient dynamic neural network model which also accounts for morphometric characteristics of the catchment. The model developed using Time-Lagged Recurrent Neural Networks (TLRNs) is used to estimate runoff for river Dikrong, a tributary of river Brahmaputra in India. Comparisons with traditional static models, with and without integration of geomorphologic data, reveal the proposed model to be a promising tool in operational hydrology.

Morphometric analysis of the submarine arc volcano Monowai (Tofua–Kermadec Arc) to decipher tectono-magmatic interactions

Morphometric analysis of multibeam bathymetry and backscatter data is applied to Monowai, a submarine volcano of the active Tofua–Kermadec Arc to map and document the structure and evolution of the volcanic centre. Low rates of erosion and sedimentation, and pervasive tectonic and magmatic processes, allow quantification through detailed structural analysis and measurement of deformation. The Slope, Aspect, Curvature, Rugosity, and Hydrology (flow) tools of ArcGIS provide a robust structural interpretation and the development of a model of Monowai evolution.A nested caldera structure with a volume of ~31km3 and a stratovolcano of ~18km3 dominate the magmatic constructs. The outer caldera is elongate along 125°, and the inner caldera along 135°. Numerous parasitic cones and fissure ridges are also observed, oriented at 039° and 041°, respectively. Northeast trending faults (with a regional average strike of 031°) are widespread within this part of the backarc, forming a nascent rift graben to the west of the Monowai caldera complex. The distribution of throw varies spatially, reaching a maximum total along-rift of 320m and across rift of 120m, with greater throw values measured in the west.Elongation directions of the two nested calderas are near-perpendicular to the trends of faults and fissure ridges. The inner caldera is more orthogonal to the magmatic constructs (fissure ridges and aligned vent cones) and the outer caldera is approximately orthogonal to the regional fault fabric, suggesting a strong interaction between magmatic and tectonic processes, and the directions of the horizontal principal stress. We present a detailed morphometric analysis of these relationships and the data are used to interpret the spatial and temporal evolution of the tectono-magmatic system at Monowai, and classify the type of rifting as transtensional. Similar analysis is possible elsewhere in the Kermadec backarc and within other regions of submarine volcanism.

A simulation of soil water content based on remote sensing in a semi-arid Mediterranean agricultural landscape

This paper shows the application of a water balance based on remote sensing that integrated a Landsat 5 series from 2009 in an area of 1,300 km2 in the Duero Basin (Spain). The objective was to simulate the daily soil water content (SWC), actual evapotranspiration, deep percolation and irrigation rates. The accuracy of the application is tested in a semi-arid Mediterranean agricultural landscape with crops over natural conditions. The results of the simulated SWC were compared against 19 in situ stations of the Soil Moisture Measurement Stations Network (REMEDHUS), in order to check the feasibility and accuracy of the application. The theoretical basis of the application was the FAO56 calculation assisted by remotely sensed imagery. The basal crop coefficient (Kcb), as well as other parameters of the calculation came from the remote reflectance of the images. This approach was implemented in the computerized tool HIDROMORE+, which integrates various spatial databases. The comparison of simulated and observed values (at different depths and different land uses) showed a good global agreement for the area (R2=0.92, RMSE=0.031 m3 m-3, and bias=-0.027 m3 m-3). The land uses better described were rainfed cereals (R2=0.86, RMSE=0.030 m3 m-3, and bias=-0.025 m3 m-3) and vineyards (R2=0.86, RMSE=0.016 m3 m-3, and bias=-0.013 m3 m-3). In general, an underestimation of the soil water content is noticed, more pronounced into the root zone than at surface layer. The final aim was to convert the application into a hydrological tool available for agricultural water management.

Estimation of soil erosion and sediment yield from mine tailing dumps using GIS: a case study at the Samgwang mine, Korea

This paper presents a Geographical Information Systems (GIS)-based method for the estimation of soil erosion and sediment yield from mine tailing dumps at abandoned mining areas. A region of 21.27 km2 in size around the Samgwang mine in Korea was selected as a study area. GIS data sets (i.e. a Digital Elevation Model (DEM), a soil map and a land cover map) and the mean annual rainfall over 30 years recorded from the nearest observatory were used to create factor layers for calculating the mean annual rate of soil erosion using the universal soil loss equation model. The watershed areas were extracted from the DEM using hydrological analysis tools in GIS. The concept of sediment delivery ratio was used to determine the sediment yield from mine tailing dumps during storm events. As a result, the soil erosion and sediment yield from mine tailing dumps in the study area were calculated as 75.63–350.24 tons yr− 1 and 40.40–187.64 tons yr− 1 respectively.

Development of hydrological tools using extreme rainfall events for Dhaka, Bangladesh

Annual daily maximum rainfall data for 57 years (1953 to 2009) in Dhaka, Bangladesh are analysed statistically to prepare two hydrological tools: (1) probable return periods of extreme rainfall events, and (2) intensity-duration-frequency (IDF) curves. Limited availability of short-duration rainfall data leads to difficulties in performing the task through traditional approaches; hence, the Gumbel distribution function and scaling theory have been applied in preparing the probable return periods of extreme rainfall events and IDF curves, respectively. The outcomes of this paper can be used in better designing hydraulic structures in Bangladesh.

Pragmatic Approaches for Water Management Under Climate Change Uncertainty1

Stakhiv, Eugene Z., 2011. Pragmatic Approaches for Water Management Under Climate Change Uncertainty. Journal of the American Water Resources Association (JAWRA) 47(6):1183–1196. DOI: 10.1111/j.1752-1688.2011.00589.x Abstract: Water resources management is in a difficult transition phase, trying to accommodate large uncertainties associated with climate change while struggling to implement a difficult set of principles and institutional changes associated with integrated water resources management. Water management is the principal medium through which projected impacts of global warming will be felt and ameliorated. Many standard hydrological practices, based on assumptions of a stationary climate, can be extended to accommodate numerous aspects of climate uncertainty. Classical engineering risk and reliability strategies developed by the water management profession to cope with contemporary climate uncertainties can also be effectively employed during this transition period, while a new family of hydrological tools and better climate change models are developed. An expansion of the concept of “robust decision making,” coupled with existing analytical tools and techniques, is the basis for a new approach advocated for planning and designing water resources infrastructure under climate uncertainty. Ultimately, it is not the tools and methods that need to be revamped as much as the suite of decision rules and evaluation principles used for project justification. They need to be aligned to be more compatible with the implications of a highly uncertain future climate trajectory, so that the hydrologic effects of that uncertainty are correctly reflected in the design of water infrastructure.

Using ensemble of climate models to evaluate future water and solutes budgets in Lake Kinneret, Israel

► Output from ensemble of four climate models were integrated into three hydrological models. ► Predictions of climate change impact on inflows, evaporation and salinity in Lake Kinneret, Israel. ► Ensemble integration prediction for water inflows (2015 -2060) is a reduction of ∼2.6 Mm 3 y -1 . ► The expected trend of evaporation (2015 -2060) is an increase of ∼0.41 Mm 3 y -1 . ► The expected trend of lake salinity (2015 -2060) is an increase of 1.18 mg Cl - L -1 y -1 . Identifying and quantifying future climate effects on water resources has major economic and societal implications, rendering such studies extremely important for water planners. Here we integrate output from one high resolution global (Japan Meteorological Agency) and three regional (ECHAM-RegCM, Hadley-MM5, ECHAM-MM5) climate models into three hydrological tools (1. annual incoming water volumes; 2. evaporation from the lake; and 3. lake salinity) to provide first approximations of climate change impacts on water quantity and quality in Lake Kinneret (also known as Sea of Galilee), the major freshwater resource in Israel. Meteorological data extracted from the climate models were used as input data into the models. Results were calculated for the historical 1979–2009 and the future 2015–2060 periods. The modeled historical period was verified against observed data, first by each model alone, and then by the combined model structure. Predicted results varied between the climate models. The ECHAM-RegCM predicted decreased precipitation in an average rate of ∼7 mm year −1 (−0.8% annually) while the trends of precipitation predicted by the other models were less obvious. According to the combination of ECHAM-RegCM, ECHAM-MM5 and Hadley-MM5 with the lake evaporation model, the evaporation will increase by 0.2–0.6 Mm 3 (0.10–0.25%) annually while according to the JMA no trend was found. The lake salinity is mostly impacted by changes in inflows and therefore only the ECHAM-RegCM predicted significant increase of salinity (from 280 ppm Cl today to ∼450 ppm Cl in 2060), while the trends of salinity according to other models were mild.

A comparison of tools for modeling freshwater ecosystem services

Interest in ecosystem services has grown tremendously among a wide range of sectors, including government agencies, NGO's and the business community. Ecosystem services entailing freshwater (e.g. flood control, the provision of hydropower, and water supply), as well as carbon storage and sequestration, have received the greatest attention in both scientific and on-the-ground applications. Given the newness of the field and the variety of tools for predicting water-based services, it is difficult to know which tools to use for different questions. There are two types of freshwater-related tools--traditional hydrologic tools and newer ecosystem services tools. Here we review two of the most prominent tools of each type and their possible applications. In particular, we compare the data requirements, ease of use, questions addressed, and interpretability of results among the models. We discuss the strengths, challenges and most appropriate applications of the different models. Traditional hydrological tools provide more detail whereas ecosystem services tools tend to be more accessible to non-experts and can provide a good general picture of these ecosystem services. We also suggest gaps in the modeling toolbox that would provide the greatest advances by improving existing tools.

Variation in curve numbers derived from plot runoff data for New South Wales (Australia)

AbstractThe curve number method is a simple one parameter (the curve number) rainfall runoff model. While its theoretical underpinning has been questioned it remains a powerful hydrological tool in the absence of detailed data and is therefore used extensively in hydrological models. This study aims to characterize the variation in maximum retention values (S), which underlie curve numbers, for a range of agricultural treatments across a large spatial area in New South Wales (NSW), Australia. The data used for the analysis spans several decades of rainfall runoff observations. A range of different derivation methods result in variation in mean and variance of S. In particular, methods that emphasize the larger storms result in greater S and thus lower runoff. For larger spatial scales, emphasis on larger storms gives more reliable estimates of S. Systematic variation in S arises from variations in treatment, pre‐runoff soil moisture, rainfall depth, and variations in cover. On the basis of the analysis, a table of curve number values for different land uses found in NSW is presented.The resulting distributions of S and curve numbers provide guidance for rainfall runoff modelling studies in the agricultural important areas of NSW. Copyright © 2011 John Wiley & Sons, Ltd.

Development of efficient and cost-effective distributed hydrological modeling tool MWEasyDHM based on open-source MapWindow GIS

Many regions are still threatened with frequent floods and water resource shortage problems in China. Consequently, the task of reproducing and predicting the hydrological process in watersheds is hard and unavoidable for reducing the risks of damage and loss. Thus, it is necessary to develop an efficient and cost-effective hydrological tool in China as many areas should be modeled. Currently, developed hydrological tools such as Mike SHE and ArcSWAT (soil and water assessment tool based on ArcGIS) show significant power in improving the precision of hydrological modeling in China by considering spatial variability both in land cover and in soil type. However, adopting developed commercial tools in such a large developing country comes at a high cost. Commercial modeling tools usually contain large numbers of formulas, complicated data formats, and many preprocessing or postprocessing steps that may make it difficult for the user to carry out simulation, thus lowering the efficiency of the modeling process. Besides, commercial hydrological models usually cannot be modified or improved to be suitable for some special hydrological conditions in China. Some other hydrological models are open source, but integrated into commercial GIS systems. Therefore, by integrating hydrological simulation code EasyDHM, a hydrological simulation tool named MWEasyDHM was developed based on open-source MapWindow GIS, the purpose of which is to establish the first open-source GIS-based distributed hydrological model tool in China by integrating modules of preprocessing, model computation, parameter estimation, result display, and analysis. MWEasyDHM provides users with a friendly manipulating MapWindow GIS interface, selectable multifunctional hydrological processing modules, and, more importantly, an efficient and cost-effective hydrological simulation tool. The general construction of MWEasyDHM consists of four major parts: (1) a general GIS module for hydrological analysis, (2) a preprocessing module for modeling inputs, (3) a model calibration module, and (4) a postprocessing module. The general GIS module for hydrological analysis is developed on the basis of totally open-source GIS software, MapWindow, which contains basic GIS functions. The preprocessing module is made up of three submodules including a DEM-based submodule for hydrological analysis, a submodule for default parameter calculation, and a submodule for the spatial interpolation of meteorological data. The calibration module contains parallel computation, real-time computation, and visualization. The postprocessing module includes model calibration and model results spatial visualization using tabular form and spatial grids. MWEasyDHM makes it possible for efficient modeling and calibration of EasyDHM, and promises further development of cost-effective applications in various watersheds.

Limitation of 90 m SRTM DEM in drainage network delineation using D8 method—a case study in flat terrain of Bangladesh

Bangladesh is the deltaic flood plain located in the lower ridge of the Ganges Brahmaputra and Meghna basins. The country is very flat having 40% of its landmass up to 10 m above the mean sea level. Shuttle Radar Topography Mission (SRTM) 90 m, i.e., 3-arc second digital elevation model (DEM) is being widely used to delineate river network and to extract catchment information using hydrology tool of ArcGIS. The hydrology tool uses the D8 method for extraction of drainage pattern. The study was carried out to find the limitation and suitability of 90 m SRTM DEM data in flat terrains especially in Bangladesh using ArcGIS. Twelve catchments of varying geomorphology were chosen from five hydrological zones of Bangladesh. Basin characteristics such as bifurcation ratio, drainage density, and channel slope of the catchments were estimated and analyzed to evaluate the suitability of 90 m SRTM DEM. The delineated catchments of slope 1:3,600 or more flat shows large deviation in river network alignment when compared with the digital river network developed by Bangladesh Water Development Board and with Google Earth’s images. The catchments having slope 1:2,850 and more steep were delineated correctly. The conclusion could not be established between slopes 1:2,850 and 1:3,600. The study also revealed that the catchment characteristics other than the slope have no effect on river network delineation.

Managing water resources in a time of global change: mountains, valleys and flood plains

Series editor's preface, Henry Vaux, Jr, 1. High Water, Low Water, No Water: Upland Water Resource Management in times of Global Change, Ariel Dinar and Alberto Garrido, Part I. Mountain and upland areas, 2. Mountains of the World - Water Towers for the 21st Century, Bruno Messerli, Daniel Viviroli and Rolf Weingartner, 3. Upland Watershed Management and Global Change - Canada's Rocky Mountains and Western Plains, Dave Sauchyn, Michael Demuth and Alain Pietroniro, 4. System Hydrology Tools To Support Water Policy: The Case Of Lake Kinneret Watershed, Israel, Alon Rimmer, 5. Combined impacts of development and climate change on the Rhone river (Switzerland, France), Jean-Paul Bravard, Part II. Managing Opportunities in the face global challenges, 6. Adaptive Management of Upland Rivers Facing Global Change: General Insights and Specific Considerations for the Rhone, Claudia Pahl-Wostl, Ger Berkamp and Katharine Cross, 7. Climate Change and Water Resources: Potential Impacts and Implications, Richard M. Adams and Dannele E. Peck, 8. Policy and Adaptation in the Jordan Valley, Yousef Hasan AYADI, 9. Upland Watershed Management in Developing World: The Case of Tunisia, Sihem Benabdallah, 10. Challenges In Upland Watershed Management: The Green-Blue Water Approach, Holger Hoff, Part III. The view of practitioners and stakeholders, 11. Factors for success: Public and Private Sector Roles in Securing a Safe and Reliable Water Supply, Don Lowry, 12. The International Joint Commission: A Case Study in the Management of International Waters, Dennis Schornack and John Nevin, 13. An overview of the IJC, Jack Blaney, 14. Water management in alpine regions, Pedro Arrojo-Agudo, Conclusions, 15. Overcoming the Constraints for a more Integrated and Adaptive Water Management: a New Agenda for Upland Waters , Alberto Garrido and Ariel Dinar.

Fractionation of oxygen and hydrogen isotopes in evaporating water

Variations in oxygen and hydrogen isotope ratios of water and ice are powerful tools in hydrology and ice core studies. These variations are controlled by both equilibrium and kinetic isotope effects during evaporation and precipitation, and for quantitative interpretation it is necessary to understand how these processes affect the isotopic composition of water and ice. Whereas the equilibrium isotope effects are reasonably well understood, there is controversy on the magnitude of the kinetic isotope effects of both oxygen and hydrogen and the ratio between them. In order to resolve this disagreement, we performed evaporation experiments into air, argon and helium over the temperature range from 10 to 70 °C. From these measurements we derived the isotope effects for vapor diffusion in gas phase ( ε diff(HD 16 O) for D/H and ε diff(H 2 18 O) for 18O/ 16O). For air, the ratio ε diff(HD 16 O) / ε diff(H 2 18 O) at 20 °C is 0.84, in very good agreement with Merlivat (1978) (0.88), but in considerable inconsistency with Cappa et al. (2003) (0.52). Our results support Merlivat’s conclusion that measured ε diff(HD 16 O) / ε diff(H 2 18 O) ratios are significantly different than ratios calculated from simplified kinetic theory of gas diffusion. On the other hand, our experiments with helium and argon suggest that this discrepancy is not due to isotope effects of molecular collision diameters. We also found, for the first time, that the ε diff(HD 16 O) / ε diff(H 2 18 O) ratio tends to increase with cooling. This new finding may have important implications to interpretations of deuterium excess (d-excess = δD − 8δ 18O) in ice core records, because as we show, the effect of temperature on d-excess is of similar magnitude to glacial interglacial variations in the cores.