In the tropics, ample freshwater is the primary resource supporting thriving human and ecological communities. In the Pacific Islands, many watersheds are threatened by climate change, urban encroachment, and invasion by water-demanding exotic plant species like strawberry guava (SG). To maintain an adequate freshwater supply, adaptive management strategies are needed to address these concerns while confronting operational barriers to implementation. We developed a prototype watershed decision support tool (WDST) that incorporated: (i) distributed hydrology modeling to quantify effects of climate change and SG invasion on freshwater yield; (ii) a decision support tool that linked potential changes in yield with treatment costs, accessibility, and conservation values to identify priority restoration and protection areas; and (iii) a collaborative process for developing, refining, and using the WDST. We estimated water yield for 904 hydrologic subunits (median=220 ac, hereafter, hydro-subunits), across a >200,000 ac study area on the windward side of Hawai’i. Water yield was calculated for current vegetation conditions and full SG invasion and removal scenarios. In the WDST’s logic model these data were used to assess opportunities for: (i) SG removal; and (ii) protection against SG invasion. A decision model was coupled with logic model outputs to incorporate budgetary constraints, access limitations, and habitat conservation value into treatment area selection. This process was repeated under 6 climate scenarios over a seven year period. Freshwater yield was most sensitive to changes in rainfall. Under the full invasion scenario, warming increased stand-level water use, but with SG absent, water yield increased due to reduced evapotranspiration by native plants. The greatest changes occurred under the warm+dry climate scenario where water yield declined by ~30%. The WDST gave high priority scores to low cost, easily accessed, high water yielding subunits, in areas with strong conservation protection. WDST results also revealed that poor access was the major factor reducing priority scores for otherwise high yielding subunits.
