Soils differ substantially in fertility and hence in their suitability for intensive agriculture - especially traditional and low-input forms of intensive agriculture. Often soil fertility and the suite of soil properties and processes that are associated with it change abruptly in space, despite gradual and continuous changes in the factors that influence soil fertility (for example rainfall and soil age). This research models the causes and evaluates the consequences of abrupt changes in soils, and explores their implications for the functioning of terrestrial ecosystems under both natural and agricultural conditions. It builds on an extensive data base of soils from the Hawaiian Islands that includes large differences in the factors that control soils (for example rainfall from < 200 mm/yr to > 4000 mm/yr, soil age from a few hundred to several million years); this data base includes very clear examples of thresholds in soil properties and processes. In addition, it draws upon a detailed analysis of the distribution and dynamics of Hawaiian agriculture prior to European contact. The modeling effort will iterate between fundamental models of soil developmental processes and this detailed information, seeking to determine how much detail is necessary to model the development of abrupt changes in soils and the regions between them where soil properties are relatively consistent. It will test those models with newly collected soils information in and around pre-contact agricultural systems in Hawaii, and on soils formed from different types of rock in New Zealand.
The project will provide fundamental information on the processes that underlie soil fertility, and how they shaped the distribution of intensive agricultural systems in the past (and could contribute to low-input systems today and in the future). Carrying out this research will require integrating information from ecology, soil science, hydrology, geochemistry, and anthropology; hence it will contribute to the conceptual and practical unification of disparate disciplines. It will support two post-doctoral fellows and several undergraduate students ? and those students and postdoctoral fellows will learn how to integrate a broad array of research fields. Finally, this project is closely aligned with ongoing efforts by Native Hawaiian organizations and institutions to understand and to restore the traditional pre-contact agricultural systems of Hawaii ? for educational and cultural purposes, and also to learn what those systems can contribute to present and future agricultural sustainability. This project will bring a scientific component to those efforts, and thus provide opportunities to teach and learn science that are integrated with the cultural and restoration efforts.
Ecosystem functioning is controlled in part by the interaction of climatic forcing with geological substrate, which produce soils. Soil properties do not change smoothly across the continents; the distribution of unique features is governed by chemical and physical responses to the forcing factors. The chemical response change in climate can be muted or can be dramatic. A minimal response occurs when the climate or other forcing factors (e.g. change in land use) creates a perturbation that is quickly buffered by chemical reactions in soil. We call these well buffered parts of the soil continuum soil process domains. By contrast, if some perturbation forces a soil process domain to the edge of the system of buffers that regulates internal soil processes, the chemical reactions that determine soil and ecosystem properties can shift dramatically. We call these shifts pedogenic thresholds. Pedogenic thresholds mark shifts in landscapes that we recognize as changes in soils and/or ecosystems. For instance, as we move from an upland to a wetland there are substantial changes in the nutrient supply, the way in which carbon compounds decompose and the types of plants that grow there (wetland plants are adapted to low oxygen conditions). As another example consider the transition from hyper-arid to arid climate conditions shown at the bottom of this figure. When rainfall is so low that no plants can grow as say in the Atacama or Sahara Deserts there is very little carbon dioxide in the soils because there is very little biological respiration. When that happens the dominant anions are cloride and sulfate and the soils become very salty. As soon as there is enough moisture to support plants the dominant anion is bicarbonate produced by solubilization of carbon dioxide and dissociation of the resulting carbonic acid. The soils then begin to accumulate calcium carbonate to form the familiar caliche deposits that are common in desert soils. The shift from sulfate to carbonate is dramatic and strongly tied to biology. Other process domains and thresholds that are responsible for the global patterns of ecosystem function. Although we understand these process domains and thresholds conceptually, we do not understand them well enough to predict how a particular ecosystem will respond to a particular perturbation, for instance climate change or transformation of forest to grassland. It was the purpose of this project to use the Hawaiian Islands and New Zealand as testing grounds to understand the relationships among climate, substrate and landscape age in determining how pedogenic thresholds are expressed on a local to regional basis. Hawaii is a really useful natural laboratory because lava flows of very similar composition are spread across the islands and have very different ages and states of soil formation. Also each island has a wet side and dry side imposed by the northeast trade winds. As a consequence we have been able to define a number of soil process domains and pedogenic thresholds there. One of the striking features we find in Hawaii is that thresholds are non-linear. The implications of this non-linearity is that an ecosystem will be more susceptible to rapid change at different combinations of age and climate. Prediction of ecosystem response is not simply a matter of knowing the magnitude of the forcing function and the resistance (resilience) of the ecosystem. One also needs to have an understanding of the buffering reactions and how much forcing they have already absorbed. Our work has been designed to develop that understanding for basalt supported ecosystems in Hawaii. However there are questions about whether we would find similar patterns of very different geological substrates. We tested that by sampling a climate gradient on post-glacial loess deposits on the south island of New Zealand. Our results there showed that similar thresholds exist although they occur at lower rainfall. In detail we interpret the difference in where thresholds occur within the context of differing chemistry on different rocks and minerals. Overall however what we find is that we can make predictions about the sensitivity of soils and ecosystems to climate and/or land use change.
