Climate and land use interact to determine phytoplankton nutrient limitation

Background/Question/Methods

A dominant paradigm assumes that phosphorus (P) is the limiting nutrient of lake phytoplankton. However, recent research suggests that nitrogen (N)-limitation exists on seasonal to decadal time frames. In human-impacted water bodies, watershed land-use explains some of the patterns in nutrient availability but climate, specifically precipitation, connects the terrestrial landscape to aquatic ecosystems. The goal of this research was to determine the interactive effects of climate and watershed land use on phytoplankton nutrient limitation status. To address this question we quantified the identity (N vs P) and severity of nutrient limitation on 400 lake-dates from reservoirs. We used the Palmer Drought Severity Index (PDSI), which couples temperature and precipitation, as an index of runoff that enters the reservoirs and calculated a principal component axis combining watershed land use (agriculture versus forest) and watershed area:lake area as our land use metric. We used multiple regression to compare all lake-dates. We also examined seasonal trends in three reservoirs, across a gradient of land use (from 80% agricultural to 80% forested) that were each sampled multiple times per season for at least 6 years. We assessed the interactive effects of climate and land use by comparing spring (wet) and summer (dry) nutrient limitation in addition to wet and dry years (estimated from  growing season PDSI).

Results/Conclusions

Using all 400 lake-dates, the model had more explanatory power in the spring (N-limitaion-22% and P-limitation-35% of variance explained) than in the summer (N-limitation-7% and P-Limitation-12%). In the spring, the severity of N-limitation was positively associated with dry periods and forested watersheds, while P-limitation was positively associated with wet periods and agricultural landscapes. Phytoplankton in the intensively-sampled reservoir in the forested watershed were consistently N-limited with only a few instances of P-limitation, while the reservoir in the agricultural watershed was P-limited in all years and seasons, except in 2007 when phytoplankton became N-limited in mid-summer. 2007 had a seasonal drought with a PDSI of -1.5 for May through October. Phytoplankton in the reservoir in the mixed (agricultural/forested) watershed was always P-limited in the spring but switched to N-limitation during the summer. The drought-associated switch to N-limitation in the agricultural reservoir was reflected in the phytoplankton community, which was dominated by non-heterocystous cyanobacteria in most summers, but switched to heterocystous cyanobacteria in 2007. Our results show that climate and watershed land use interactively determine phytoplankton nutrient limitation status  and species composition.