Improving the Performance of an Advanced Land Surface Model for Simulating Hydrology of Ecuadorian Amazon subbasins

The Amazon region is considered a biodiversity hotspot and one of the most important regions in terms of water availability and an essential dynamic element of the macroclimate system. The Amazon catchment is composed of 19 major subbasins that span seven countries. In Ecuador, there are three main subbasins: the Napo, Pastaza, and Putumayo. Flow in all three is affected by meltwaters of tropical glaciers in the Andes and by rainforest micro-climate interactions. Yet, the hydrological characteristics of these subbasins are difficult to represent in physical-based models because of the complexity of the terrain, lack of observations, and localized climatic variability. For instance, the satellite and reanalysis datasets misinterpret the transition from bimodal precipitation (Ecuadorian Andes) to unimodal in the Amazon rainforest, resulting in miscalculations of monthly discharges. To address these challenges, it is important to characterize multiple sources of uncertainty and to merge data streams from models, satellites, and available in situ observations. Here, we apply the NASA Land Information System (LIS) software framework to simulations of the Napo subbasin, in an effort to evaluate the contribution of different input data sources and models to reliable simulation of hydrological regimes. We present LIS results of the Noah-MP 4.0.1 Land Surface Model (LSM) and the HyMAP2 routing model. The simulations cover two subbasins within the Napo, the first one in the highlands paramo dominant ecosystem and the second one in the lowlands of the Amazon rainforest. We find compensating errors, as the simulations perform better in the lowland subbasin while showing errors in the highland subbasin. This system is being refined in order to provide a reliable basis for water-based decision tools. This understanding would in turn help to ensure water availability and sustainable use of water resources in the region.