Davide Saurwein | University of Zurich, Switzerland (original) (raw)

Uploads

Papers by Davide Saurwein

Hydropower and water resources management are an important issue in most countries of the world, ... more Hydropower and water resources management are an important issue in most countries of the world, included Switzerland. The scientific community is currently engaged to produce numerical models and simulations which aim at understanding the most important concurring factors of climate with the support of different tools and methods, at both regional and global scales. In this case, the purpose is to simplify reality while reducing errors and uncertainties related to streamflow prediction. The latter uncertainties can be due to different sources, such as the initial hydrologic conditions of a catchment, the hydrological model’s input data and structure, or a too high amount of subjectivity which is applied while implementing such modelling procedures.

The aim of this work is to investigate the propagation of uncertainties from the input meteorological forecasts to the resulting streamflow predictions. A weather generator has been used to create synthetic weather decadal forecasts. These forecasts have then been fed into the hydrological model Hydrologiska Byråns Vattenbalansavdelning (HBV), and simulations have been run in order to obtain corresponding runoff forecasts. The accuracy of the meteorological and runoff forecasts has been calculated with similar statistical metrics with the aim to assess the propagation of uncertainty. Three statistical metrics, defined as “skill scores”, have been applied for this purpose. The experiment was performed for two glacierized catchments both located in the Swiss Alps, Findelen and Gries. The simulations were performed by assuming different scenarios of glacier extent in order to observe the influence of the amount of ice present in the catchment on the results. The effect of a varying input glacier extent on simulated runoff has then been studied, together with an assessment of the modifications on the hydrological regime. The simulations have been run by applying the recently-implemented glacier routine in the hydrological model with different settings, with the aim to analyze how skill transfer can be affected. In addition, a sensitivity analysis has been performed on the parameters and routines of the hydrological model in order to study their contribution to model efficiency.

It has been observed that the influence of precipitation on runoff forecasts is lower than the one of temperature for highly-glacierized catchments. This influence increases with diminishing glacierization. In a hypothetical ice-free catchment, the effect of precipitation on skill transfer tends to become more relevant, for both Findelen and Gries catchments. Another important factor of skill transfer is the lead time from which a forecast is produced, and also the morphological and topographical features of the catchment. Moreover, the application of different settings of the glacier routine can also have an influence on skill transfer and on the influence of temperatures and precipitation.

Keywords: Decadal forecast, streamflow forecast, synthetic forecast, forecast skill, forecast lead time, uncertainty propagation, skill score, scenario, “perfect” model, glacierized catchment, Hydrologiska Byråns Vattenbalansavdelning (HBV) model.

Climate is nowadays a main topic of research between the scientific community, because of its com... more Climate is nowadays a main topic of research between the scientific community, because of its complexity and because of the influence of a lot of parameters in order to determine the evolution of climate in the middle and long term. Climate is concerned with the description of middle-and long term variability of variables like temperature and precipitations (but also irradiation, wind and relative humidity), and it should not be confused with the meteorology, which is the study of these variables at one specific date for a short period of time (usually one or several days). This thesis presents an analysis of data from the Swiss Permafrost Monitoring Network (« PERMOS »), the national meteorological institute of Switzerland and the projects « TEMPS » and « ENSEMBLES ». The analysis is done for the period between 1961 and 2099 for 12 mountain stations at middle and high altitudes in Switzerland. This analysis of climate has the objective to do a detailed description of the variability of eight climate variables: relative humidity, precipitations, irradiation, air temperature (mean, minimum and maximum), mean speed of the wind and maximum speed of the wind. The analysis of the evolution of the variables is done for the whole period 1961-2099, for the past (1961-2015), for the future (2016-2099), and also for the norm periods of 20-30 years The analysis is done following the two-step approach presented by RAJCZAK ET AL. (2015) in their scientific article. The approach proposes an analysis of data of climate variables for the past in order to produce a possible scenario for the evolution of these variables into the future (especially for the next century). Observations of the national meteorological institute of Switzerland and of the Swiss Permafrost Monitoring Network are integrated with the analysis of RAJCZAK ET AL. (2015) in order to propose a scenario for the future. Therefore, data from the meteorological stations of the national institute of meteorology and climatology are integrated with data from the 12 stations described in this thesis (analysed firstly by RAJCZAK ET AL., 2015). The objective is to propose a climate scenario which is based upon the models « GCM »

Hydropower and water resources management are an important issue in most countries of the world, ... more Hydropower and water resources management are an important issue in most countries of the world, included Switzerland. The scientific community is currently engaged to produce numerical models and simulations which aim at understanding the most important concurring factors of climate with the support of different tools and methods, at both regional and global scales. In this case, the purpose is to simplify reality while reducing errors and uncertainties related to streamflow prediction. The latter uncertainties can be due to different sources, such as the initial hydrologic conditions of a catchment, the hydrological model’s input data and structure, or a too high amount of subjectivity which is applied while implementing such modelling procedures.

The aim of this work is to investigate the propagation of uncertainties from the input meteorological forecasts to the resulting streamflow predictions. A weather generator has been used to create synthetic weather decadal forecasts. These forecasts have then been fed into the hydrological model Hydrologiska Byråns Vattenbalansavdelning (HBV), and simulations have been run in order to obtain corresponding runoff forecasts. The accuracy of the meteorological and runoff forecasts has been calculated with similar statistical metrics with the aim to assess the propagation of uncertainty. Three statistical metrics, defined as “skill scores”, have been applied for this purpose. The experiment was performed for two glacierized catchments both located in the Swiss Alps, Findelen and Gries. The simulations were performed by assuming different scenarios of glacier extent in order to observe the influence of the amount of ice present in the catchment on the results. The effect of a varying input glacier extent on simulated runoff has then been studied, together with an assessment of the modifications on the hydrological regime. The simulations have been run by applying the recently-implemented glacier routine in the hydrological model with different settings, with the aim to analyze how skill transfer can be affected. In addition, a sensitivity analysis has been performed on the parameters and routines of the hydrological model in order to study their contribution to model efficiency.

It has been observed that the influence of precipitation on runoff forecasts is lower than the one of temperature for highly-glacierized catchments. This influence increases with diminishing glacierization. In a hypothetical ice-free catchment, the effect of precipitation on skill transfer tends to become more relevant, for both Findelen and Gries catchments. Another important factor of skill transfer is the lead time from which a forecast is produced, and also the morphological and topographical features of the catchment. Moreover, the application of different settings of the glacier routine can also have an influence on skill transfer and on the influence of temperatures and precipitation.

Keywords: Decadal forecast, streamflow forecast, synthetic forecast, forecast skill, forecast lead time, uncertainty propagation, skill score, scenario, “perfect” model, glacierized catchment, Hydrologiska Byråns Vattenbalansavdelning (HBV) model.

Climate is nowadays a main topic of research between the scientific community, because of its com... more Climate is nowadays a main topic of research between the scientific community, because of its complexity and because of the influence of a lot of parameters in order to determine the evolution of climate in the middle and long term. Climate is concerned with the description of middle-and long term variability of variables like temperature and precipitations (but also irradiation, wind and relative humidity), and it should not be confused with the meteorology, which is the study of these variables at one specific date for a short period of time (usually one or several days). This thesis presents an analysis of data from the Swiss Permafrost Monitoring Network (« PERMOS »), the national meteorological institute of Switzerland and the projects « TEMPS » and « ENSEMBLES ». The analysis is done for the period between 1961 and 2099 for 12 mountain stations at middle and high altitudes in Switzerland. This analysis of climate has the objective to do a detailed description of the variability of eight climate variables: relative humidity, precipitations, irradiation, air temperature (mean, minimum and maximum), mean speed of the wind and maximum speed of the wind. The analysis of the evolution of the variables is done for the whole period 1961-2099, for the past (1961-2015), for the future (2016-2099), and also for the norm periods of 20-30 years The analysis is done following the two-step approach presented by RAJCZAK ET AL. (2015) in their scientific article. The approach proposes an analysis of data of climate variables for the past in order to produce a possible scenario for the evolution of these variables into the future (especially for the next century). Observations of the national meteorological institute of Switzerland and of the Swiss Permafrost Monitoring Network are integrated with the analysis of RAJCZAK ET AL. (2015) in order to propose a scenario for the future. Therefore, data from the meteorological stations of the national institute of meteorology and climatology are integrated with data from the 12 stations described in this thesis (analysed firstly by RAJCZAK ET AL., 2015). The objective is to propose a climate scenario which is based upon the models « GCM »