Water balance in paired watersheds with eucalyptus and degraded grassland in Pampa biome

Abstract Rangelands of the Pampa biome, which cover regions of Argentina, Uruguay and Brazil (176,496 km 2 – 2.07% of Brazilian territory and 63% of Rio Grande do Sul State territory, southern region of Brazil) in South America (total area of 750,000 km 2 ), are being substituted by crops and commercial eucalyptus, with potential impacts on ecological and hydrological response of watersheds and river basins. We evaluated the influence of vegetation cover on hydrological processes by describing the water balance and its components (rainfall – P, interception, throughfall, actual evapotranspiration – ETa, groundwater recharge – G, and streamflow – Q) in two paired watersheds. One watershed being cropped to 7-year-old Eucalyptus saligna stands (forest watershed – FW; 0.83 km 2 ) and another consisted of degraded grassland with native and exotic grasslands (grassland watershed – GW; 1.10 km 2 ) used for livestock production in the Rio Grande do Sul State. The study was conducted from October 2012 to September 2014, during two hydrological years: a normal year from October 2012 to September 2013, and a wetter year from October 2013 to September 2014. Pluviometers were installed to study the partition of rain, along with watershed gauges and equipment for monitoring hydrological variables. Meteorological data (maximum and minimum temperature, relative humidity, wind speed, and solar radiation) used to calculate potential evapotranspiration were collected from a tower installed in the FW, whereas hydrological data (P and Q) were collected by sensors installed in each of the watershed spillways. During the normal year, P was 19% above the annual historical average for the region, which is 1314 mm, whereas in wetter year, rainfall was 98% above the same average. Total rainfall interception was similar between years in GW (9 and 10%), but different between years in FW that was higher in wetter (24%) then in normal year (16%). In the normal year, streamflow were 64% lower in the FW compared to the GW, while ETa and G were respectively 37% and 25% greater in the FW compared to the GW. In the wetter year, streamflow was 66% lower in the FW than in the GW, while ETa and G in soil were respectively 27% and 46% greater in the FW compared to the GW. Flow with 5% time streamflow (Q 5 ) was greater in the GW compared to FW in both normal and wetter years. Streamflow in the GW and FW were equal at Q 80 and Q 82 in the normal and in the wetter years, respectively, and exceedance probability curves crossed over at Q 81 and Q 82 , where the exceedance probability curves become greater in FW than in the GW. Even if the forest watershed had greater ETa compared to the grassland watershed, benefits such as greater interception and lesser surface runoff can be highlighted, for a condition where grassland was degraded and provided low aboveground biomass. Thus, the cultivation of eucalyptus stands may provide better structural conditions and ground cover, greater infiltration and soil water retention, and increased groundwater recharge, with consequent reduction of soil degradation by erosion and increased water availability during dry periods. Long-term use of forest systems, especially when compared to degraded grassland, may provide improvement on soil physical quality. However, these comparative results may not be valid for conditions during harvesting and tillage operations, neither when soil under grasslands has improved physico-hydraulic properties.