br Conclusions br Acknowledgements This


This work was supported by the Vermont Experimental Program for Stimulating Competitive Research (EPSCoR) Award number NSF EPS Grant # 1101317. The authors acknowledge the Vermont Advanced Computing Core which is supported by NASA (NNX 06AC88G), at the University of Vermont for providing High Performance Computing resources that have contributed to the research results reported within this paper. We acknowledge the World Climate Research Programme\’s Working Group on Coupled Modelling, which is responsible for CMIP, and we thank the climate modeling groups (listed in Table A-1 of this paper) for producing and making available their model output. For CMIP, the U.S. Department of Energy\’s Program for Climate Model Diagnosis and Intercomparison provides coordinating support and led development of software infrastructure in partnership with the Global Organization for Earth System Science Portals. The authors are indebted to two anonymous reviewers for their constructive and thorough reviews.

Water use by deep-rooted forests is in some cases higher than shallow-rooted agricultural crops or pasture (Zhang et al., 2001). Prior to European settlement, the Atlantic Forest region of Brazil was predominantly covered by rainforest occupying an area of approximately 150Mha distributed across large latitude and climatic ranges (Ribeiro et al., 2009). By the late 1900s, however, clearing for agriculture and population growth left only about 11–16% coverage of this native forest (Instituto National de Pesquisas Espaciais, 2011; Ribeiro et al., 2009). During the past four decades, plantation forestry has expanded rapidly in parts of this region mainly by replacing pasture. Eucalypt plantations in Brazil now occupy more than 5.1Mha with a significant portion in this Kinase Inhibitor Library (ABRAF, 2012). This plantation expansion has largely occurred on the plateau part of the landscape away from streams, while rainforest covers areas closest to the streams (gullies). These large-scale changes in vegetation may have affected stream flows, but these effects have not been widely quantified. Despite the mean annual precipitation of around 1200mm, stream flow is intermittent and low water availability during the dry season probably limits plantation growth (Stape et al., 2008).
Plantation management, particularly harvesting, and the proportional area of and water use by the native forest, could affect stream flow. However, the potential effect of harvesting the native forest has not been measured in this landscape. Such an experiment was not possible as harvesting of streamside native forest is prohibited by the Brazilian forest code (Presidência da República, Lei N° 12.651, 2012), in which it is referred to as the area of permanent preservation (APP).
Generally, stream flow is affected by precipitation amount, distribution and intensity, soil water holding capacity and soil hydraulic conductivity, vegetation interception and water use, and terrain topography. Water use by vegetation depends on plant water demand and actual uptake, which in-turn depends on factors including rooting density and depth, available soil water and climate. The hydrological effects of native forest fragmentation are poorly understood, especially in the tropics in relation to horizontal interactions between land units of vegetation, topography and soils (Giambelluca, 2002). Previous studies in Espírito Santo State of the Atlantic Forest region have quantified many of the aspects of the hydrological cycle in one dimension (plot scale) at plantation and native forest sites, in which water balance simulations generally matched observations (Soares and Almeida, 2001; Almeida and Soares, 2003; Almeida et al., 2007, 2010). A similar capability has become available in two dimensions (e.g. hillslope) using the HYDRUS model (Smethurst et al., 2013), which enabled extension of water balance simulations at the same site to the headwater catchment scale and thereby take into account both vegetation types, i.e. Eucalyptus plantation and native forest. Our objectives here were to (1) determine if the HYDRUS model could be used to adequately simulate the hydrology of a headwater catchment in the Atlantic Forest region of Brazil used for plantation forestry with native forest in the APP, and if so (2) use it to understand the potential impacts of harvesting of one or both vegetation types on ground water and stream flow. Our hypothesis was that a quantitative analysis using the HYDRUS model would demonstrate that water use by the native forest APP could prevent an increase in stream flow after harvesting up-slope eucalypt plantations.