Determination of sample size for estimating ion throughfall deposition under a mixed hardwood forest at the Lake Clair Watershed (Duchesnay, Quebec)

Variability of throughfall volume and ion deposition were measured under a mixed hardwood forest using 36 or 72 collectors from 1988 to 1994 during the growing period (May to October) for 163 events. The ions showing the highest variability in results were (in decreasing order) NH4+, H+, Mg2+, K+, Ca2+, Cl-, Na+, NO3-, and SO42-, based on the median coefficient of variation of deposition. The variability observed was related to the reactivity of the ion in the canopy, with the more reactive ions showing higher spatial variability. The number of collectors required to obtain estimates of throughfall ion depositions within a 10% error and a confidence interval of 95% varied from 21 (SO42-) to 165 (NH4+) on a weekly basis based on events of median variability. The error associated with deposition measurements was decreased considerably from weekly, to monthly, to seasonal deposition (3 months). The number of collectors necessary to measure throughfall deposition depends on the data quality objectives of the study (i.e., error and confidence interval chosen as acceptable), the variability specific to a given ion, and the time period needed to study a given process with adequate temporal resolution. A figure is proposed that may be used as a guideline to determine the number of collectors needed in mixed hardwood forests based on the latter information.