Extrinsic and intrinsic forcing of regime shifts: A 3,000-year record of climate and lake-productivity changes in Minnesota
Regime shifts represent large, rapid, and long-lasting changes to ecosystems. Theory predicts that abrupt biological regime shifts may occur in response to large or gradual environmental forcings (extrinsic and intrinsic regime shifts, respectively). Despite advances in detecting regime shifts in controlled experimental studies, distinguishing between extrinsic and intrinsic regime shifts in field settings remains a challenge. If ecological processes are resistant to large climate forcing or if such changes are infrequent, then intrinsic shifts may dominate. Alternatively, if ecological processes are more sensitive to abrupt than gradual changes in climate, then extrinsic shifts may dominate. To assess these hypotheses we conducted sediment analyses at a dimictic lake in Minnesota to infer centennial-to-millennial scale changes in lake productivity and climate during the past 3,000 years. Lake productivity was inferred from biogenic silica (BSi), δ13C of calcite, and varve thickness data. Aridity was deduced from δ18O of calcite from the same lake. Change-point analyses were used to assess regime shifts in the time series of each indicator and the extent to which regime shifts in aquatic indicators were synchronous with shifts in aridity.
On centennial scales, BSi, varve thickness, and δ13C display concurrent regime shifts toward increased aquatic productivity at ~1600 and 250 years BP, whereas other change points in these indicators were asynchronous. No centennial-scale changes in aridity occurred before 1600 years BP, as inferred from the δ18O record, but mean δ18O values sharply decreased at ~1600 and 250 years BP and centennial-scale variability of δ18O changed at ~1200, 800, and 600 years BP. Together, these results indicate an overall trend toward greater aquatic productivity and wetter conditions during the past 3,000 years, probably because more nutrients wash into the lake and/or mixing is increased when conditions are wetter. Synchronous shifts in all indicators at ~1600 and 250 years BP suggest that decreased aridity led to regime shifts in aquatic productivity. However, changes in the variability of aridity from 1200-600 years BP did not coincide with shifts in aquatic indicators, and regime shifts in lake productivity occurred before 1600 years BP when δ18O was relatively stable. These results indicate that aquatic productivity is sensitive to large environmental changes as well as local factors related to intrinsic shifts, and they suggest that certain ecosystems, such as in our study, may withstand changes in long-term environmental variability to greater extent than mean conditions.