ICP-Forests (International Co-operative Programme on Assessment and Monitoring of Air Pollution Effects on Forests): Quality Assurance procedure in plant diversity monitoring
2009
Allegrini, Maria-Cristina | Canullo, Roberto | Campetella, Giandiego
Knowledge of accuracy and precision rates is particularly important for long-term studies. Vegetation assessments include many sources of error related to overlooking and misidentification, that are usually influenced by some factors, such as cover estimate subjectivity, observer biased species lists and experience of the botanist. The vegetation assessment protocol adopted in the Italian forest monitoring programme (CONECOFOR) contains a Quality Assurance programme. The paper presents the different phases of QA, separates the 5 main critical points of the whole protocol as sources of random or systematic errors. Examples of Measurement Quality Objectives (MQOs) expressed as Data Quality Limits (DQLs) are given for vascular plant cover estimates, in order to establish the reproducibility of the data. Quality control activities were used to determine the “distance” between the surveyor teams and the control team. Selected data were acquired during the training and inter-calibration courses. In particular, an index of average cover by species groups was used to evaluate the random error (CV 4%) as the dispersion around the “true values” of the control team. The systematic error in the evaluation of species composition, caused by overlooking or misidentification of species, was calculated following the pseudo-turnover rate; detailed species censuses on smaller sampling units were accepted as the pseudo-turnover which always fell below the 25% established threshold; species density scores recorded at community level (100 m² surface) rarely exceeded that limit.
显示更多 [+] 显示较少 [-]