Subproject 2 will use field data, remote sensing data and downscaled projections from the Subproject 1 climate model to quantify green and blue water availability and pasture productivity to test (1) if blue water availability is a function of watershed size, slope, and land cover, and if dry season green water depends on precipitation, soil texture, time since deforestation, and stocking rate, (2) if deforestation in a watershed improves the availability of surface water for the properties in that watershed in both the wet and dry seasons, (3) if intensification, as represented by increased stocking density, reduces the ability of water to seep into soil, resulting in more overland flow in the wet season, (4) if soil water availability at the end of the dry season decreases as sand content, stocking density, and time since deforestation increase, and (5) if soil water availability correlates with remotely-sensed non-photosynthetic vegetation.
The streamflow data for the estimation of dry season blue water will come from existing data, and from field measurements in small watersheds during the dry season. The areal coverage of ponds will be determined using visual interpretation of satellite imagery for small ponds and with a product created using Mapbiomas and Landsat TM imagery for larger ponds. Infiltration capacities and soil properties will be measured with a portable rainfall simulator. Soil moisture and meteorological data will be collected at two pasture sites, and will be used to calibrate and validate the downscaled land surface model (SSiB2).
Methods and Activities
Property-level green water availability will be calculated using the SSiB model down-scaled to the property level. Blue water availability, including both discharge of streams that drain through the property and the volume of on-farm storage will be determined, including the mean minimum discharge over 5, 10, 30, and 60 day periods. Each index will be calculated using regional regression equations. Remote sensing data will be used to document pasture condition and test for its correlation with modelled soil moisture from the downscaled land surface model.