Regionalisation of Hydrological Indices to Assess Land-Use Change Impacts in the Tropical Andes

Abstract:

Andean ecosystems are major water sources for cities and communities located in the Tropical Andes; however, there is a considerable lack of knowledge about their hydrology. Two problems are especially important: (i) the lack of monitoring to assess the impacts of historical land-use and cover change and degradation (LUCCD) at catchment scale, and (ii) the high variability in climatic and hydrological conditions that complicate the evaluation of land management practices. This study analyses how a reliable LUCCD impacts assessment can be performed in an environment of high variability combined with data-scarcity and low-quality records. We use data from participatory hydrological monitoring activities in 20 catchments distributed along the tropical Andes. A set of 46 hydrological indices is calculated and regionalized by relating them to 42 physical catchment properties. Principal Component Analysis (PCA) is performed to maximise available data while minimising redundancy in the sets of variables. Hydrological model parameters are constrained by estimated indices, and different behavioural predictions are assembled to provide a generalised response on which we assess LUCCD impacts. Results from this methodology show that the attributed effects of LUCCD in pair-wise catchment comparisons may be overstated or hidden by different sources of uncertainty, including measurement inaccuracies and model structural errors. We propose extrapolation and evaluation in ungauged catchments as a way to regionalize LUCCD predictions and to provide statistically significant conclusions in the Andean region. These estimations may deliver reliable knowledge to evaluate the hydrological impact of different watershed management practices.