Meeting:Abstract 2011 CSDMS meeting-009
Modeling the Glacial-Interglacial Impact of the Pacific Trade Wind Inversion on the Geomorphology and Hydrology of the Big Island of Hawaii
[[Image:|300px|right|link=File:]]The interaction of the subsiding, subtropical limb of the Hadley circulation and the easterly North Pacific trade winds establishes a persistent thermal inversion about halfway up the eastern flank of the Big Island of Hawaii. This restricts convective rainfall to the lower elevations, resulting in stream channels that cross an order-of-magnitude rainfall gradient, active ephemerally above the inversion and perennially below it. Above the inversion–capped cloud layer, precipitation is on the order of 400 mm/yr, and the landscape features thin, weakly-developed soils, gentle hillslopes, and ephemeral, shallowly incised bedrock streams and grassland gullies. Below the inversion, where rainfall is >3000 mm/yr, the perennial streams run through 50- to 100-m-deep gulches, with steep forested walls covered by thick tropical soils that are prone to landsliding. Meter- to 50-meter waterfalls are common downstream of the inversion layer, and incision of the deep gulches may proceed by upstream migration of these knickpoints from the coast. The positions of these knickpoints likely reflect the history of lava flows in these catchments, base level changes due to landsliding at the coast, and the statistics of water and sediment discharge above and below the trade inversion and through time.
This landscape has evolved entirely in the last 0.3 Ma, and thus under conditions of glacial-interglacial climate oscillations. During glacial periods, the inversion’s average elevation was likely depressed, although the magnitude of this depression is not well-constrained. An ice cap that was present on Mauna Kea altered the hydrology of the upper slopes of the mountain, providing a continuous source of meltwater to channels that, in the modern setting, are active only during winter storms and rare hurricane strikes. The frequency and intensity of such storms during glaciations are also not well-known.
To quantify these effects, we would like to use climate models to inform landscape evolution models. A key difficulty in coupling these types of models is the separation of time and spatial scales involved. Global climate models typically run on grids of 1 degree or more, at temporal resolution of seconds and run lengths of years to decades. Landscape evolution models (LEMs) reside at the other end of both dimensions, with typical spatial resolutions of meters to km and temporal resolutions of years or decades. The entire duration of a climate model run may be shorter than the timestep of a typical LEM.
We report initial results from our efforts to bridge the relevant scales by downscaling large-scale climate model output for last-glacial and modern times with NCAR’s regional-scale Weather Research and Forecasting (WRF) model. The predicted precipitation fields are input to a hydrologic model to generate realistic discharge statistics useful for landscape modeling. This modeling chain may be validated for the modern climate using atmospheric observations, including the modern distribution of inversion height, and USGS stream gauge data. For glacial periods, the ability of the weather model to correctly predict snowlines on Mauna Kea provides a first-order point of calibration.