Grass Control Plus Phosphorus Fertilization Result in Reduced Soil Water Content in a Montane Acacia Koa Forest in Hawaii.

Reforestation of abandoned pastures with koa (Acacia koa, Gray), an important endemic tree species, rarely eliminates shade-tolerant introduced grasses, which may compete with the koa trees and inhibit regeneration of other native forest species. A split-plot study was established in 2001 in a 23 year-old koa forest on the island of Hawaii to develop a mechanistic understanding of koa responses to silviculture. The ~1,700-m elevation site receives ~2,000 mm mean annual precipitation and is located on 1,500 to 3,000 year-old shallow Hapludands. The main treatment was thinning; sub-treatments were grass control (GC), and grass control plus phosphorus (P) fertilisation (GC+P). Volumetric soil water content was continuously monitored in the top 20-cm of soil in unthinned plots during 2005-2007 using Decagon’s ECH₂O sensors, which uses capacitance to measure soil dielectric permittivity. On an annual basis, grass control alone had no effect; however the CC+P mean 24 hour water content (0.130 m³ m^-3) was significantly less (p<0.0001) than the control. (0.167 m³ m^-3). On a weekly basis, the level of precipitation determined the level of the treatment effect. Additional measurements were obtained with the portable Delta-T ThetaProbe, which works on similar principles as the ECH₂O sensor, also showed significantly lower (p<0.0001) volumetric water content in the GC+P treatment compared to the control. These results suggests that soil water use by koa increased in response to P fertilisation, possibly due to increased koa leaf area or greater photosynthesis. Grass control alone had little effect. Despite concerns of the accuracy of using capacitance to determine soil water content, it appeared that the two technologies were sensitive enough to measure treatment effects in a forest soil.