Paul, James T. 1996. Long-term cumulative effects and water quality: A trend analysis examining the impacts of forest management and natural variability on three watersheds on the Olympic Peninsula. M.S.
Long-term cumulative effects of forest management on water quality have become a greater concern in recent years, yet few studies have been conducted that address this issue. Three large watersheds (100s of km2) on the Olympic Peninsula in Washington state, the Hoh, North Fork Quinalt, and North Fork Skokomish, are examined in this study. The Hoh watershed has experienced significant forest harvesting activities while the NF Quinalt and NF Skokomish are managed exclusively for recreation. The seasonal Kendall trend analysis is used to analyze nitrogen, phosphorus, dissolved oxygen, and stream temperature over periods ranging from 8 to 22 years. The results are compared to available records of land-use history and natural variability.
A significant increase in levels of forest harvesting on the Hoh watershed corresponds to upward trend in nitrogen concentrations between 1972 and 1980. However, it is unclear if this is a cause-and-effect relationship or a result of natural variability. An upward trend in phosphorus levels during the dry season (April-September) from 1973-1980 corresponds to an upward trend in air temperatures. It is possible that an increase in the rate of melting of four permanent glaciers feeding the Hoh may be the cause via an increase in sediment delivery to the stream. No trends exist for any of the constituents between 1973-85 on North Fork Quinalt (1976-1986) and are the result of natural processes. There is evidence that there may be some sort of long-term natural cycles in average stream temperatures occurring within the NF Quinalt.
Natural variation among large-scale watersheds (100s of km2) may be as significant, if not more significant, than human-caused variation. Differences in the results between the two unmanaged watersheds (NF Quinalt and NF Skokomish) are as significant as the differences between the managed and unmanaged watersheds (NF Quinalt and Hoh). Only in the case of the nitrogen trend on the Hoh is the explanation possibly not to do with natural variability. In terms of the monitoring water quality parameters in large watersheds, it is difficult to distinguish between trends caused by moderate land-management activities and those caused by natural spatial and temporal variation. Diffuse, non-point sources of changes in water quality regimes by human manipulations of large-scale watersheds may be indistinguishable from the effects of natural, long-term processes already occurring within the watershed.