Diatom biomonitoring of streams: Reliability of reference sites and the response of metrics to environmental variations across temporal scales

Abstract

Benthic diatoms are widely used indicators of human impacts on stream ecosystems because they are very responsive to changing environmental conditions. However, little research has explicitly focused on their reliability with regards to temporal variation in assemblage structure and environmental conditions. We examined variability in diatom-environment relationships at bi-weekly, monthly, and yearly time scales from 7 reference, 7 agricultural, and 2 acid mine drainage (AMD)-impacted streams, and how nutrient and pH fluctuations may affect the interpretation of diatom metrics and the Diatom Model Affinity (DMA) index. Reference streams had less bi-weekly variability in NO 3-N concentrations than non-reference streams. The % eutraphentic diatoms and DMA scores were more strongly correlated with seasonal means of NO 3-N and PO 4-P concentrations than with same day concentrations. Most nutrient indicator metrics had strong correlations with watershed land use. All 14 non-AMD streams experienced substantial increases in NO 3-N and decreases in temperature from November to May, which were associated with high species turnover, substantial changes in community structure, reduced diversity and richness, increased relative abundances of high nutrient diatoms, and decreases in low nutrient diatoms and DMA scores. The % acidophilic diatoms and DMA scores were significantly correlated with increased pH associated with greater precipitation at AMD sites from December to April ( r = −0.77, r = 0.62, respectively; P < 0.01). Yearly, DMA scores for all reference streams were consistently in the minimally impaired category, whereas scores for non-reference streams varied among impairment categories. Reference sites serve as reliable benchmarks for diatom ecological integrity during the summer. In this region, June to October is a recommended time period for diatom sampling in monitoring programs because subsequent shifts in hydrologic regimes, nutrients, and diatom assemblages occurred, affecting all sites and masking among stream differences attributable to agricultural land uses.