Seasonal patterns of evapotranspiration from cleared and uncleared tropical savanna: implications for catchment water balance in the wet-dry tropics.

Vegetation forms a key component of surface hydrological processes and a thorough understanding of surface processes must include a knowledge of the patterns of vegetation water use, both seasonally and spatially. This is particularly important given the extremes in rainfall of the seasonal tropics of north Australia. A major investigation is underway to examine spatial and temporal patterns of evapotranspiration (ET) from contrasting land uses in north Australia, focussing on the Daly River catchment in the NT. Within the Daly River catchment, eddy covariance towers have been installed at sites representative of uncleared savanna, recently cleared savanna with native pastures and long term cleared with improved pastures. Data has been collected since September 2007 across this tower network and collection is ongoing, This initial data set will provide ET from these contrasting land types from the late dry (September) and through the wet season of 2008, a period when 75% of the annual ET occurs. ET from uncleared savanna vegetation (LAI ~ 1) in the late dry season was approximately 2 mm d-1. By contrast, ET at the improved pasture site was at 0.5 mm d-1. Uncleared savanna vegetation used more water during this late dry season period, due to the presence of evergreen and deeply rooted (5 m) Eucalypt trees. By the mid-wet season (January), ET from the pasture site had increased by an order of magnitude and was at near potential evapotranspiration rates (6 mm d-1) and higher than the uncleared savanna site. Continuing data will enable us to see how long these high rates persists at the cleared sites. Previous work suggests that at the grass-dominated cleared sites, ET decreases dramatically with the onset of the dry season, whereas savanna vegetation maintains rates of water use through the dry season, exploiting deep soil horizons. Contrasting root architecture is reflected in levels of soil moisture of at 1 m depth, which is significantly higher at the cleared site, with far less extraction at depth occurring when compared to uncleared, deep rooted woody savannas. Changes in soil moisture and drainage have occurred rapidly (within 10 years) and suggests wide-spread clearing will have an impact on surface water-groundwater interaction, with potentially significant increases in drainage. Measurements of surface processes (ET, soil moisture dynamics) will be complimented by data describing recharge under these sites of differing times since clearing. Estimates of sub-catchment and catchment scale ET will be undertaken using remote sensing and models, which will be calibrated using the tower-derived ET measurements.