R. Kirnbauer

Distributed Snowmelt Simulations in an Alpine Catchment: 1. Model Evaluation on the Basis of Snow Cover Patterns

This paper presents an attempt at deterministically modeling spatially distributed snowmelt in an alpine catchment. The basin is 9.4 km2 in area and elevations range from 1900 to 3050 m above sea level. The model makes use of digital terrain data with 25 m grid spacing. Energy balance components are calculated for each grid element taking topographic variations of solar radiation into account. For each grid element albedo and snow surface temperatures are simulated. Model performance is evaluated on the basis of snow cover depletion patterns as derived from weekly air photographs. The use of spatially distributed data allows for addressing individual model components. Results indicate that the basic model assumptions are realistic. Model inadequacies are shown to arise from processes not included in the model such as avalanching and long wave emission from surrounding terrain as well as inaccurate model parameters.

Distributed Snowmelt Simulations in an Alpine Catchment 2. Parameter Study and Model Predictions

A distributed grid-based model is used (1) to analyze the importance of selected model parameters, (2) to simulate spatial distributions of snow cover properties in a small basin and (3) for a comparison with less sophisticated models as typically used in operational applications. Results indicate that variations of water equivalent with slope and local relief are of utmost importance for realistic distributed simulations but more moderately influence mean basin melt. Snow cover variables of which spatial distributions are simulated include the thermal and hydraulic state of the pack and hourly melt water release. All variables exhibit substantial variations in space and time. They are primarily controlled by topography and the delay of melt water in deep packs. The grid model is compared with a snow band model and a parametric model. The latter estimates the snowpacks areal extent from water equivalent. Simulated snow-covered areas suggest the grid model to be the most realistic. Differences in terms of mean basin melt derive from different assumptions associated with model structure.

Point snowmelt models with different degrees of complexity — Internal processes

Abstract Two point snowmelt models are compared under different weather and snowpack conditions. The research-oriented simulation model describes the coupled heat and mass flow distributed with depth whereas the operational model treats the snowpack as one piece and uses heat and liquid water storage factors. An identical approach to energy input is used in both cases. Parameters are derived externally from the literature. The distributed model gives satisfactory results for all periods analysed. The performance of the bulk model depends on snowpack conditions. Particularly during freeze-thaw cycles, proper calibration of parameters appears to be essential. An analysis of the state variables cold content and liquid water content indicates that the bulk model is not suited to a detailed simulation of internal processes.

Modelling snowmelt in a mountainous river basin on an event basis

Abstract A snowmelt model for short time flood forecasting of mixed rain-snowmelt floods in a high alpine watershed has been developed. The model is based on a subdivision of the basin into elevation bands. The energy input into the snowpack is computed by the energy-balance approach, using physically based preset parameters. The internal processes are parameterized by introducing heat and water storage capacities. The state of the snow cover throughout the basin is charactacterized by distinguishing three zones with different melt and drainage conditions. The lowest zone is saturated and runoff-producing. Above it, there is a transition zone of partly soaked snow. In the uppermost portion of the basin, no liquid water is stored in the snow. Snow line, saturation line and dry-snow line form the boundaries between the respective zones. Time variations of snow cover conditions are described by the altitudinal fluctuations of the lines. In performing simulation runs, the three boundary lines are found to follow different patterns during a six-day test period. Due to the prevailing melt conditions, the snow line rises monotonically and is only slightly influenced by different weather conditions. The saturation line and consequently the band width of the soaked zone, however, are controlled by day-to-day and diurnal changes in meteorological variables and exhibit a significant increase on rainy days and pronounced fluctuations during fair weather. The dry-snow line shows minor fluctuations. A sensitivity analysis indicates that the influence of model parameters on simulated melt rates is moderate or small when simulation periods of several days are considered so that parameters may be pre-set without inducing much additional uncertainty. The influence of model parameters on simulated melt and dynamics of basin snow cover conditions is discussed. The snowmelt routine was developed with the intention of starting it during the ablation period. Thus, initial conditions for the above mentioned boundary lines are required. Based on sensitivity analyses, it is found that the elevation of the snowline must be derived from current observations. For saturation line and dry-snow line, simple relations to air temperature are given. Simulation results indicate that the areal extent of the saturated snow cover must be considered, if proper model performance for the first hours after model start is desired.

Identifying Space-time Patterns of Runoff Generation: A Case Study from the Lohnersbach Catchment, Austrian Alps

Runoff generation is a result of the interplay of a range of processes, the relative magnitudes of which vary, among other things, with climate, catchment properties, and catchment scale. The variability of runoff generation processes within a mountain catchment and the variability from event to event is one particularly intriguing aspect. A better understanding of these spatio-temporal patterns of runoff generation is critical for obtaining realistic model simulations of events, such as extreme floods, and of run-off behaviour associated with changes in environmental and land use conditions. Estimating runoff generation is very difficult as it involves a high degree of extrapolation. Difficulties in accurately assessing runoff in mountains have been highlighted by local-scale field experiments (e.g. Scherrer 1997), observations in experimental basins (e.g. Anderson et al. 1997; Kirnbauer and Haas 1998; Torres et al. 1998; Muller and Peschke 2000; Uchida et al. 2001), and modelling studies (e.g. Moore and Grayson 1991) that emphasize the spatially highly heterogeneous nature of runoff. Also, different runoff processes may dominate at different spatial scales (see e.g. Bloschl 1996; Uhlenbrook and Leibundgut 1997). Although it is possible to estimate runoff for yet unobserved situations with hydrological simulation models, the reliability of such estimates is notoriously poor, particularly when moving from the plot scale or small catchment scale to medium sized catchments (DFG 1995). There is still a gap between the understanding of runoff generation processes at the plot scale and process-based hydrological modelling at the catchment scale.

GIS-gestützte Ausweisung von hydrologischen Umsatzräumen und Prozessen im Löhnersbach-Einzugsgebiet (Nördliche Grauwackenzone, Salzburger Land)

KurzfassungAm Beispiel des mesoskaligen Einzugsgebietes des Löhnersbaches (Kitzbüheler Alpen, ca. 16 km2) wird erläutert, wie auf der Basis von ExpertInnenwissen zur Landschaftsgenese und allgemein verfügbarer Daten mit Hilfe eines Geographischen Informationssystems (GIS) hydrologische Homogenbereiche (Hydrotope) rasterbasiert ausgewiesen werden können. Es wurde ein Regionalisierungsansatz für die flächenhafte Abschätzung der Struktur- und Lithovarianz des quartären Hangschutts entwickelt, welcher auf der Hangentwicklungsgenese basiert. Jeder der ausgewiesenen hydrologischen Homogenbereiche ist sowohl durch die gleichen prozessrelevanten Umsatzräume, als auch durch die gleichen dominanten Abflussbildungsprozesse gekennzeichnet. Die Ausweisung erfolgt prozessebenenspezifisch, so dass Prozesse der Geländeoberfläche, Zwischenabflussprozesse der Deckschichten, sowie basisabflussbildende Prozesse der Kluftgrundwasserleiter unterschieden werden. Die Plausibilität der für das mesoskalige Löhnersbach-Einzugsgebiet erzielten hydrologischen Raumgliederung konnte auf Basis einer empirischen Raumgliederungskarte geprüft werden.SummaryA method to delineate hydrological response units (HRU) within a GIS environment using generally available data sets and expert knowledge on landscape genesis is presented for the meso-scale Löhnersbach basin, located in the Austrian alps (Kitzbühler Alpen, 16 km2). Based on the genesis of the hillslope material a regionalization approach was developed which delineates the spatial structure and the lithological variance of the quaternary drift covers. Each of the delineated HRUs is characterized by the same runoff source areas and the same dominating runoff generation processes. The delineation was conducted using different process levels, differentiating between near surface processes, interflow processes in the drift cover, and base flow generating processes in the fractured hard rock aquifer. The delineation approach applied to the mesoscale Löhnersbach basin and the plausibility was proofed by a comparison with an empirically generated map.

Hydrological modeling in alpine catchments: sensing the critical parameters towards an efficient model calibration

For the Tyrolean part of the river Inn, a hybrid model for flood forecast has been set up and is currently in its test phase. The system is a hybrid system which comprises of a hydraulic 1D model for the river Inn, and the hydrological models HQsim (Rainfall-runoff-discharge model) and the snow and ice melt model SES for modeling the rainfall runoff form non-glaciated and glaciated tributary catchment respectively. Within this paper the focus is put on the hydrological modeling of the totally 49 connected non-glaciated catchments realized with the software HQsim. In the course of model calibration, the identification of the most sensitive parameters is important aiming at an efficient calibration procedure. The indicators used for explaining the parameter sensitivities were chosen specifically for the purpose of flood forecasting. Finally five model parameters could be identified as being sensitive for model calibration when aiming for a well calibrated model for flood conditions. In addition two parameters were identified which are sensitive in situations where the snow line plays an important role.

Flood Forecasting for the River Inn

The river Inn as the main river in Tyrol moulds the settlement and economic area in Northern Tyrol in a considerable way. 66 % of the area drains into the Inn, whereas the remaining 34% drain into the Lech, the Grossache and the Drau in East Tyrol. The Inn flows through Tyrol for about 200 km, from the Swiss border at Martinsbruck to Kufstein, where it leaves Tyrol and flows into Bavaria/Germany (Fig. 2.1).

The Kühtai data set: 25 years of lysimetric, snow pillow, and meteorological measurements

Abstract Snow measurements at the Kühtai station in Tirol, Austria, (1920 m.a.s.l.) are described. The data set includes snow water equivalent from a 10 m2 snow pillow, snow melt outflow from a 10 m2 snow lysimeter placed at the same location as the pillow, meteorological data (precipitation, incoming shortwave radiation, reflected shortwave radiation, air temperature, relative air humidity, and wind speed), and other data (snow depths, snow temperatures at seven heights) from the period October 1990 to May 2015. All data have been quality checked, and gaps in the meteorological data have been filled in. The data set is unique in that all data are available at a temporal resolution of 15 min over a period of 25 years with minimal changes in the experimental setup. The data set can therefore be used to analyze snow pack processes over a long‐time period, including their extremes and long‐term changes, in an Alpine climate. Analyses may benefit from the combined measurement of snow water equivalent, lysimeter outflow, and precipitation at a wind‐sheltered alpine site. An example use of data shows the temporal variability of daily and 1 April snow water equivalent observed at the Kühtai site. The results indicate that the snow water equivalent maximum varies between 200 and more than 500 mm w.e., but there is no statistically significant temporal trend in the period 1990–2015.

Bayesian Estimation of Design Floods under Regional and Subjective Prior Information

Usually design floods are estimated as certain quantiles of a cumulative distribution function (CDF) fitted to a sample of yearly maxima of floods observed at a gauging station. If such observations do not exist at a site where a hydraulic structure is to the built, the hydrologist can collect flood data for several years during the planning phase. This small new sample, however, will not be sufficient for estimating the design flood by means of common flood statistics but it fits as one source of information to be combined with some prior information within the BAYESIAN estimation procedure. Prior information can be taken from long time flood records observed at stations of the same region to be incorporated in a data-based a-priori density function of the parameters.