Does the commonly used estimator of nutrient resorption in tree foliage actually measure what it claims to?

A descriptive temporal model is considered to be the best available estimator for accretion, resorption and proportional nutrient resorption. However, ecological studies rarely collect sufficient data for applying such a model. A less-demanding and commonly used estimator for proportional resorption (PR) calculates PR as the percentage of the nutrient pool that is withdrawn from mature foliage before leaf abscission. Data from an intensive sampling campaign of the aboveground nutrient pools and fluxes of two Betula pendula Roth. stands were used. We showed that the commonly used estimator is not an accurate estimator for accretion, resorption and proportional resorption. The commonly used estimator underestimated the proportional resorption of N on the average by 3-10%, and the proportional resorption of P by 20-25%. The low accuracy of the estimations was shown to be caused by a lack of selectiveness of the commonly used estimator. In other words, the commonly used estimator does not measure the underlying processes in specific nutrient accretion and resorption at the stand level. However, when a sufficiently high sampling density with several samples at a given point in time is used, then the commonly used estimator preserves the ranking relationship between the PR of different sites for N in 97% of the cases and for P in 71%. The commonly used estimator can thus be used in comparative studies as an index for proportional nutrient resorption only. The quantitative results should not be taken literally, as they are based on only two sets of observations. However, the results show that the commonly used estimator should no longer be used as a measure for accretion, resorption or PR whenever the plant accretes nutrients in the foliage as a compensation for nutrient losses due to foliar leaching and litterfall during the growing season.