Climate, and more specifically rainfall, is amongst the most important forces that shape a landscape, yet quantifying the impacts of rainfall on landscape development remains a challenge. It is easy to see the differences between the topography in arid versus humid landscapes, but it is has proven difficult to quantify how physical processes such as soil development, surface and subsurface water flow, and hillslope and fluvial erosion, differ between these two environments. The Kohala Peninsula, on the northern tip of the Big Island of Hawaii, offers an excellent field environment to address this question. Across the approximately 20 km wide peninsula, rainfall rates very from upwards of 4,000 mm/year to less than 200 mm/year. Deeply incised gulches dominate the topography of the wet side of the peninsula, while on the dry side, shallow gulches disappear and reappear across the landscape. The wet side of the peninsula has developed deep soils and the process of weathering is so intense that a fingernail can scrape a groove in many rocks. In contrast, on the dry side, surface soils are often shallow and rocky and exposed rock surfaces are much less weathered. This project will quantify the differences in hydrologic and erosion processes on the contrasting sides on the peninsula using both field monitoring and numerical modeling. Two watersheds, one on the dry side and one on the wet side, will be equipped with devices to measure rainfall rates, water flow depths on both the hillslopes and in the gulches, soil moisture content, and erosion rates. These devices will operate continually throughout the duration of the project. Field surveys will provide detailed observations of the geometry of the gulches and the variables that control incision rates, such as sediment grain size, rock hardness, and degree of fracturing. These data will be incorporated into a numerical landscape evolution model. The model will be used to systematically explore feedbacks between key climatic controls and surface process responses over the time scales at which landscapes evolve. Field data will be used to both constrain physical processes in the model and inspire the specific modeling scenarios.
This project will address a fundamental scientific question of how rainfall patterns actually influence the processes of weathering and erosion. Rainfall gradients are present across almost every mountain range on Earth and the findings from this study will have broad significance beyond the Hawaiian Islands. Knowledge of the detailed interactions among climate, hydrology, and erosion can be applied to landscapes across the planet. Furthermore, by quantifying links between landscape development and rainfall, the effects of potential climate change on landscape evolution can be predicted more effectively. This project will develop a number of numerical models that will be freely available for use throughout the scientific community, so that the findings from this study can be easily applied in other settings and to a wide range of scientific questions.
This project explored the impact of rainfall gradients on landscape evolution over the past ~500,000 years on the northern tip of the Big Island of Hawaii, an area known as the Kohala Peninsula. Even from Google Earth (see figure 1) the extreme rainfall gradient in this setting is apparent; the wet side is green and the dry side is brown. Some areas on the wet, northeastern side of the peninsula receive as much as 4 meters (13 feet) of rainfall per year, whereas on the dry, southwestern side of the peninsula, there are often years in which some locations receive no rain at all. The Kohala peninsula was chosen to study to impacts of rainfall on landscape evolution because many of the other factors that affect landscape evolution over long time scales, such as variable rock type or complicated patterns in rock uplift are not present in this setting. The entire peninsula is volcanic rocks, specfically basalt. The island is slowly sinking, and it is thought to be sinking relatively uniformly. The Earth Sciences community has often debated the relative role of climate and tectonics in driving landscape evolution, however, it can be difficult to decipher the influence of the two factors in a single location. The Kohala peninsula is ideal because patterns of rock uplift do not affect the region. We focused primarily on the evolution of the rivers, or gulches as they are known locally. The gulches are incising into bedrock and have numerous waterfalls, although often these waterfalls are dry. We observed that the river valleys were deeper and the waterfalls were higher on the wet side of the peninsula, indicating that river incision rates are greater on the wet side of the peninsula (see figure 2). We related the valley depth to local rainfall rates and our initial hypothesis was that there was a threshold rainfall rate needed to trigger significant fluvial bedrock incision. We tested this hypothesis using a one-dimensional computer model of river evolution. We found that the conversion of rainfall into surface water flow that then drives river incision, which is the proposed mechanism that links rainfall and river incision rates, could not account for the observed morphology of these gulches. However, if we assume that the local rock weathering rate is related to the local rainfall rate, and more weathered rocks are easier to erode, then we are able to generate river profiles similar to those observed in Kohala. Interestingly, our initial proposed threshold rainfall rate to trigger significant fluvial incision is similar to a rainfall threshold needed for significant weathering that was observed in another study. In other words, there is some evidence that weathering and bedrock incision are linked, and our modeling suggests this link could be in weakening the rocks enough so that river incision is possible. Ongoing work will test this hypothesis with data from the field area. Based on these results, we explored the potential impacts of rainfall gradients in more general settings, in order to guide research in other settings that is looking for the imprint of rainfall on landscape evolution and landscape form. To do this, we linked a two-dimensional computer model of rainfall generation that responds to the shape of the topography with a two-dimensional computer model of fluvial bedrock incision. We found that when landscapes and rainfall patterns evolve together, the shape of river networks may be quantitatively different from what would expected if the landscape evolved with uniform rainfall (see figure 3). This adjustment in network shape can make the imprint of rainfall gradients on river profile morphology much more subtle than has been predicted in previous studies using one-dimensional river profile models that do not explicitely account for network shape. These results highlight the integrated response of the entire landscape, and suggest that future studies must take a broad view when linking landscape morphology with rainfall patterns. This project has had impacts on graduate education, and both basic and applied scientific research. The results have implications for long-term landscape evolution and the building and shaping of mountainous topography. The results also have implications for understanding how rainfall, and rainfall changes, could affect river erosion in the short-term. Finally a number of students took part in this project. One PhD student was funded by this project, and he has completed his degree. Two other graduate students from Tulane University also participated in field work related to this project. The Tulane team has led on seven conference presentations and two peer reviewed published papers to date. Two more manuscripts are in preparation from the Tulane team based on this project. Finally, all of the computer code written for this project is available for other scientists to use.
