Linking stream nitrogen fluxes to multi-scale temporal patterns of precipitation

Background/Question/Methods

Around most of the world precipitation is becoming less-frequent but more-intense. The drying and rewetting cycles caused by patterns of precipitation change are likely to affect the mechanisms of nitrogen retention in watersheds by reducing plant uptake and microbial immobilization while temporarily increasing nitrification rates. Because the timing of intense rainfall events is more likely to align with less effective nitrogen retention mechanisms, watershed should lose more nitrogen when subjected to less-frequent, more-intense precipitation patterns. The hypotheses of this research were (1) wavelet methods expose meaningful relationships in real-world ecological data, and (2) the lag between precipitation and nitrate flux is smallest for low-frequency, high-intensity rainfall. I used long-term records of precipitation, stream discharge and nitrate concentrations to analyze the relationship between precipitation and stream nitrate flux using wavelet decomposition, cross-wavelet transforms and wavelet coherence. Time-series were analyzed both as raw nitrate flux and as nitrate flux corrected against chloride flux. The analyzed streams are located across an urban to rural gradient in the Baltimore area. This work was made possible by data generously provided by the Baltimore Ecosystem Study LTER under National Science Foundation Grant No. 0423476, and by NSF Grant No. DGE-0504195.

Results/Conclusions

I found little or no synchronization between precipitation and stream nitrate flux in the two forested reference watersheds at annual time scales (no wavelet coefficients within 95% confidence interval). However, precipitation and nitrate flux in urban and suburban watersheds were significantly synchronized at the annual time-scale. Furthermore, nitrate flux consistently lagged behind precipitation in these watersheds. Road salting contributes to chloride ion flux in these watersheds and may contribute to some of the pattern found at annual scales. In the smallest forested watershed nitrogen flux and precipitation are not synchronized at time scales of less than a year. The differences between forested and suburban/urban watersheds could be a consequence of differences in impervious surface area and drainage mechanisms. Despite the low temporal resolution of stream concentration records, wavelet analyses appear to be useful for identifying temporal scales of synchronization.