Effects of long-term water table drawdown on evapotranspiration and vegetation in an arid region phreatophyte community

Abstract

Evapotranspiration rates and the ground water component of evapotranspiration at a site in Colorado's San Luis Valley that is dominated by shrubby phreatophytes (greasewood and rabbitbrush) were compared before and after a water table drawdown. Evapotranspiration (ET) rates at the site were first measured in 1985–1987 (pre-drawdown) when the mean water table depth was 0.92 m. Regional ground water pumping has since lowered the water table by 1.58 m, to a mean of 2.50 m. We measured ET at the same site in 1999–2003 (post-drawdown), and assessed physical and biological factors affecting the response of ET to water table drawdown. Vegetation changed markedly from the pre-drawdown to the post-drawdown period as phreatophytic shrubs invaded former wetland areas, and wetland grasses and grass-like species decreased. Lowering the water table reduced estimated total annual ET from a mean of 409.0 to 278.0 mm, a decrease of 32%, and the ground water component of ET (ET g), from a mean of 226.6 to 86.5 mm, a decrease of 62%. Two water table depth/ET models that have been used in the San Luis Valley overestimated the reduction in ET g due to lowering the water table by as much as 253%. While our results corroborate the generally observed negative correlation between ET rates and water table depth, they demonstrate that specific models to estimate ET as a function of water table depth, if not verified, may be prone to large errors. Both the water table drawdown and the vegetation change are continuing 20 years after the drawdown began, and it is unclear how site ET rates and processes will differ after the water table has stabilized and vegetation has adjusted to the new site hydrologic conditions.