Chapter 12 Synthesis: Reading Ecosystems as Systems

Strong ecological explanations are sequences. Nutrient input does not simply cause fish death: it relaxes producer limitation, increases a bloom, increases shading and organic substrate, raises decomposer respiration, lowers dissolved oxygen and then excludes oxygen-dependent animals. Predator removal does not simply reduce biodiversity: prey or herbivores increase, their resources decline, habitat structure changes and dependent species respond. Each arrow names a testable process.

Keep the currencies separate. Energy is transferred and dissipates; carbon and nutrients change form and cycle; organisms move and reproduce; pollutant concentration can accumulate. An arrow may show more than one currency, but the explanation should identify which. Likewise, distinguish a stock such as standing biomass or atmospheric carbon from a rate such as productivity, respiration or emission.

First identify the response variable, units, spatial scale and time interval. Describe the pattern with direction, magnitude, threshold, lag or anomaly and quote data when available. Then propose a mechanism linked to chapter processes. Separate observation from explanation. A graph showing lower abundance near a shore demonstrates a distribution pattern; salinity intolerance is a hypothesis until salinity is measured and alternatives are controlled.

Next evaluate the design. Ask how samples were located, whether independent replicates exist, whether detection probability changed, and whether the method's assumptions hold. For capture–recapture, inspect closure, mixing and recapture count. For a mesocosm, inspect controls and wall effects. For a time series, distinguish trend from seasonality and autocorrelation. For a conservation intervention, demand an appropriate baseline, comparison and enough time to reveal delayed effects.

Finally make a proportionate conclusion. A statistically significant association is evidence against a specified null model, not proof of one biological cause. A model prediction is conditional on assumptions. A useful evaluation states what the evidence supports, what remains uncertain and which next measurement or manipulation would discriminate between explanations. That is the central practice of ecology: extracting causal understanding from complex, variable living systems without pretending the complexity is absent.