Trend detection in seasonal data: from hydrology to water resources

- We propose a tool (MASH) for Exploratory Data Analysis and trend detection in climate, hydrology and water resources.
- MASH can effectively handle time series with seasonality and interannual variability.
- Significant hydrological trends do not necessarily reflect into water resources.

SummaryIn this paper we investigate the relationship between hydro-climatic trends and their impacts on water resources at the basin scale, focusing on a catchment on the Italian and Swiss Alps in the period 1974-2010. More generally, we address the topic of trend detection in environmental time series combining novel and traditional tools in order to simultaneously tackle the issue of seasonality and interannual variability, which usually characterize natural processes. The paper's contribution is twofold. First, we propose a novel tool to be applied in Exploratory Data Analysis, named MASH (Moving Average over Shifting Horizon). It allows to simultaneously investigate the seasonality in the data and filter out the effects of interannual variability, thus facilitating trend detection. We describe how to combine the MASH with statistical trend detection tests, like the Mann-Kendall test, the Seasonal Kendall test, and the Linear Regression test, and Sen's method, to quantify the trends occurring in different seasons. Second, we estimate the impacts of hydrological changes in terms of water resources and we discuss their relevance from the water resources management perspective. We define and simulate a set of indicators of performances, resilience, reliability, and vulnerability, so to assess the ability of the water resources systems to absorb changes in the hydrological patterns. The analysis reveals that, in the case study area, statistically significant trends in hydro-climatic records have been undergoing in the last decades, although they have had limited impacts on water resources.