FIGURE 1.1 Comparison of natural erosion rates (over geological time) to agricultural soil erosion rates in relation to rates of soil production. The graph line comprising squares shows the rates of natural soil production, the circles show natural geological erosion, and the top line of diamonds shows agricultural erosion far exceeding the other two rates. SOURCE: Montgomery (2007a).

practical and scientific reasons why we need to better understand the impacts of humans on Earth’s physical environment.

Because natural processes vary spatially and across scales, a geographical perspective is essential to understanding their nature and character. The perspectives and tools of the geographical sciences used by geographers, geologists, ecologists, and others provide insights into soil erosion, flood magnitude and frequency, and ecological adjustments to climate change on both contemporary and paleotimescales. One significant area of investigation focuses on watershed response to and recovery from environmental changes, including Quaternary (past 2-3 million years) climatic changes and historical human-induced landscape changes. For example, because river systems respond to the integrative effects of climate and watershed processes, changes in streamflow, channel properties, and fluvial deposits provide information on the timing, direction, and magnitude of postglacial climate changes, suggesting that even modest climate shifts can generate significant changes in streamflow (Knox, 1993). Analyses of fluvial stratigraphic records have proved to be important because the identification of paleoflood occurrence extends the researchable time frame of these low-frequency events well beyond the stream gauge record, thus improving flood forecasting (Enzel et al., 1993; Baker, 1998) and capturing the periodicity of highly variable climatic episodes such as El Niño events (Gomez et al., 2004; Magilligan et al., 2008). These paleorecords suggest that climatic stationarity (the mean and variance of a time series) has not remained constant over time (Milly et al., 2008), which raises questions about existing water allocation arrangements because the stationarity assumption is the cornerstone of dam design and water allocation strategies. Higher resolution and longer-term datasets, such as those that can come from dendrochronology, can help capture these statistical shifts.

The geographical sciences have contributed to our understanding of floods as well, especially in relation to land-use changes. The massive construction of dams over the past several hundred years has had a profound impact on the hydrological regime (Figure 1.2), often leading to hydrological modifications exceeding the impacts of climate change (Magilligan et al., 2003; Magilligan and Nislow, 2005). Using archival national data, Graf (1999) identified more than 80,000 dams that have been constructed in the United States—essentially 1 dam per day on average since the signing of the Declaration of Independence. Graf ’s examination of the geographical location and context of these dams showed marked regional variations in dam number and type; most of the dams in the United States are in the eastern half of the country, although dams with the greatest impact on storage are found in the West (Graf, 1999, 2001). This pattern suggests that, although watershed fragmentation may be considerable in the eastern United States, ecological impacts due to flow reductions may be more significant in the western part of the country. Other field-based studies have provided fundamental insights into the profound