EFFECTS OF SURFICIAL GEOLOGY, LAKES, AND LAND USE ON URBAN STREAM TEMPERATURES: METROPOLITAN SEATTLE

Stream temperatures result from complex interactions of net radiation, hyporheic exchange, sensible heat exchange, and groundwater and surface water inflows. Many of these factors are influenced by urbanization, but their relative importance is obscure in urban watersheds, despite their important ecosystem effects for species such as cold-water fish. The purpose of this study was to evaluate the relative importance of factors potentially controlling summer stream temperatures using a regional dataset of near-instantaneous temperature measurements from western Washington collected during a single 2-hour period in August in each of the years 1998–2001. Each year included at least 500 measurements; the peak year (1999) covered almost 800 sites. Multiple independent site visits confirmed a measurement precision about ±1^oC, much less than the range of measured temperatures. Stream temperature was analyzed as a function of GIS-determined watershed area, percent development, percent glacial outwash deposits in the contributing watershed, and percent lake area within the drainage area of each sampling point. An information-theoretic approach (AIC) was used to determine which of nine candidate models best predicted the variability of stream temperature departures from the daily average temperature. This allowed comparison of all years in the same analysis by controlling for interannual variation in absolute stream temperatures. When examining all years of data, the analysis suggests that the only good candidate model was the global model (w_i = 0.99), which incorporated all four parameters, while all other candidate models had weights of evidence that were too low to be considered. For data only from year 1999, the candidate model using only ‘percent outwash’ had the highest weight (w_i = 0.4), followed by a combination of percent development + percent outwash, percentage of lake area only, the global model, and the model that considers three of the four parameters (percent development + watershed area + percent outwash). This analysis suggests that urbanization is not the most important factor controlling stream temperature in this region, and that geology, the presence of lakes, and watershed area have strong effects on stream temperature.