Abstract
In Norway, threats of geohazards are high, one of these geohazards that causes great damage every year is the springflood. As the climate is changing, it is expected that there will be significant changes to the trend, variability, and seasonality of precipitation, as well as snow/rain ratio of the precipitation. These changes in combination with an increased temperature will change the hydrological regime in Norway and will highly affect the yearly springfloods. The aim of being able to foresee and forecast the yearly floods relies on peoples understanding of the hydrological regime and hydroclimatic drivers of the floods. In this study the climatic drivers to yearly springflood have been investigated with regards to their relationship and effect on the yearly springflood. Twenty-five catchments in central and eastern Norway have been studied to identify, and if possible, quantify, how the interaction between the climatic variables contributes to controlling the yearly springflood. Two main methods were used to carry out this study. Correlation analysis was used to identify the relationship between the climatic variables and the flood characteristics. Multiple linear regression approach was used to better understand the interaction between the climatic variables and the yearly springflood. With this approach regression models were developed to use in predicting future springflood events with the climatic variables as input. Results suggest that trends in the flood characteristics of the springfloods are highly depended on the climatic drivers, catchment properties and in which flood regime the catchments fit into. The regression models that were developed to predict when the springflood hits were mostly dependent on the snow cover in the catchments, the groundwater table and the temperature prior to the springflood. The regression models that were developed to predict the peak and volume (sum of the maximum seven-day discharge) of the springflood were mostly dependent on the amount of snow, number of frost days, snow-cover, snow-water equivalent and groundwater table in the catchments. When evaluating the performance of the developed regression models, the results showed that the models developed to predict the timing and volume of the springflood gave the best results. The trends in how the twenty-five studied catchments were affected by a change in precipitation patterns and snow/rain ratio due to climate change were also investigated. Negative trends in the springfloods peak and volume were found more often than positive change. This can be explained by a shift in flood generating processes as there is expected to be a shift in the snow/rain ratio in precipitation in Norway. The results of this thesis suggest that the role of snowmelt as a flood generating factor for springfloods tends to decrease, with rain becoming a more dominant factor in generating the yearly springflood.