How to minimize the sampling effort for obtaining reliable estimates of diel and annual CO2budgets in lichens

Abstract

AbstractEstimating carbon budgets for poikilohydric organisms, such as lichens and bryophytes, requires methods other than those for homoiohydric plants due to a strong dependency of carbon gain on fluctuating hydration. This paper provides guidance with respect to optimal sampling strategies for estimating annual carbon budgets of lichens and bryophytes, based on a one-year dataset of half-hourly CO2-exchange readings on the epilithic placodioid lichenLecanora muralis(syn.Protoparmeliopsis muralis) and tests the effects of reduced sampling frequencies and different temporal sampling schemes on carbon budget estimates. Both fine-scale sampling (measurements within a day) and large-scale sampling (selection of days within a year) are addressed.Lowering the sampling frequency within a day caused large deviations for 24-h (diel) budget estimates. Averaged over a larger number of days, these errors did not necessarily cause a large deviation in the annual budget estimate. However, the occurrence of extreme deviations in diel budgets could strongly offset the annual budget estimate. To avoid this problem, frequent sampling (c. every 1·5 hours) is necessary for estimating annual budgets. For estimating diel budgets and patterns a more frequent sampling (everyc. 0·5 hours, balancing data resolution and disturbance) is often needed.Sampling fewer than 365 days in a given year inevitably caused estimates to deviate from the ‘true’ carbon budget, i.e. the annual budget based on half-hourly measurements during 365 days. Accuracy increased with total sample frequency, and blocking days caused larger deviations than sampling randomly or regularly spaced single days. Restricting sampling to only one season led to strongly biased estimates. The sampling effort required for a reliable estimate of the annual carbon balance of lichens based on simple extrapolations of diel carbon budgets is impracticably large. For example, a relatively large sample of 52 random days yielded an estimate within 25% of the true annual budget with only 60% certainty. Supporting approaches are therefore suggested, in particular extrapolating diel budgets using ‘weather response types’, possibly aided by diel activity patterns from chlorophyll fluorescence, or modelling CO2exchange as a function of climatic conditions.