Estimating root longevity at sites with long periods of low root mortality

Minirhizotron technique is capable of providing median root longevity. The use of the median longevity might overestimate root longevity if the distribution of survival times is very skewed or irregular, as is the case at sites where root mortality is very low during the long winter. In this paper we illustrate the case theoretically and compare that with field observation in northern Sweden to show an alternative procedure for such sites. Hypothetical root cohorts were constructed to investigate and show some technical problems with estimating median root longevity at a Swedish northern site where root mortality is very low during long winter time (8 months), and to investigate whether these problems could be overcome by discarding winter time from the survival analysis and include only the growing season in which the roots are at risk of mortality. Authentic root data, gathered in a minirhizotron study at such a site, were analysed on a whole year basis and on season basis. By analysing longevity based only on the season when there is a risk for root death, the median longevity became a more reliable estimate of the true mean longevity. When this method was applied to root data from northern Sweden, the estimated root longevity in different treatments became between 17% lower and 8% higher compared to the longevity estimated on a whole year basis.We conclude that the reliability of the median longevity as an estimate of the true mean longevity can be increased by basing the survival analysis only on the parts of the year when fine roots are at risk of mortality at sites with long winter and low root mortality.