Extending observations further: using historic biogeochemical data to understand trends in Puget Sound

Fluctuations in Puget Sound water quality reflect a combination of natural variability and anthropogenic influence. Predictions of how future changes will unfold requires an understanding of the complex interplay between these factors. Recent and ongoing measurement of water quality parameters allow a thorough assessment of variability over short time scales, but understanding changes over longer (e.g., decadal) timescales is critical for forecasting future changes. Examining long term trends requires the use of historical data, however, measurements that pre-date modern monitoring programs can be scarce, and the detection limits are often higher than newer analytical techniques. Here, we will consider several possible methods to stitch together modern and historical data records of dissolved oxygen and nutrient concentrations while accounting for differences in sample rate, temporal coverage, and sensitivity of analytical techniques. To test different data analysis methods, data records with good temporal coverage that show discernible trends and/or high covariance with other environmental parameters will be used as the basis for simulating data with spotty temporal coverage (as might be expected for less commonly measured biogeochemical parameters like deep dissolved oxygen). Selected trend analyses will then be applied to these simulated time series records to assess method effectiveness in identifying a known feature. Top performing methods for trend analysis will then be applied to key biogeochemical parameters such as nitrate and deep dissolved oxygen. We report on the major covariates and trends identified and the associated statistical confidence intervals. These results will be part of a larger effort to determine how existing trends might play out under different anthropogenic input scenarios