Distribution Of Fine Root Longevity For Manchurian Ash And Davurian Larch In Plantation Forest

Root systems, especially fine roots, which link the above- and belowground ecosystem processes, have become one of the most concerned focuses in ecology research. The turnover of fine roots plays an important role in ecosystem carbon allocation and nutrient cycles. Unfortunately, accurately quantifying below-ground C fluxes and estimating fine root longevity are still elusive. One of reasons is because we do not know the proportions of shortlived vs. long-lived root pools or the distribution of fine root longevity within these pools. To examine these issues in detail, the distributions of fine root longevity were studied in two plantations, an ash (Fraxinus mandshurica) and a larch (Larix gmelinii) in Northeast China, where fine root diameter, order, and the effects of fertilization on fine root were measured over a two-year period by using minirhizotrons. The main results were showed as follows:(1) The whole distributions of Ash and larch fine roots longevity consistently showed skewed bimodal distributions, which were composed of two distinct components. The main component was short-lived root pool with large kurtosis and narrow distribution range, containing the whole dead fine roots in one growing season. The other component had a wide distribution range and smaller kurtosis. Root longevity, analyzed by growing-season cohort, was characterized by each of the four theoretical probability distributions: Normal, Welibull, Lognormal, and Exponential. The results of testing for fit distribution showed that Welibull, Normal, and Lognormal were good, and there was not significance difference between them. On the other hand, the Exponential distribution can not pass statistical test mostly. The whole fine roots longevity distribution can be expressed with a sum of two Weibull, Normal, and Lognormal distributions that represented two components of fine root longevity. By fitting measured fine root longevity data to proposed distributions, the best theoretical probability distribution for Ash was Welibull, mean while, for Larch was Lognormal.(2) Fine roots＜0.5 mm in diameter accounted for most of the total number of individual roots sampled in both Ash and Larch. So the Frequency of fine roots longevity distribution for Ash and Larch were positively skewed relativing to traditional definition of fine roots. In regard to one growing season, the result of testing for distribution goodness of fit showed that Welibull was the best one in most diameter classes. On the other hand, for two growing season the Lognormal distribution could best characterize the longevity distribution in numerous diameter classes, the Welibull and Normal taked the second place.(3) Although both the total amounts and dead number of Ash root were obviously more than Larch ones, the number of explicit former two orders of fine roots for Ash was less than Larch one. Whether diameter or longevity, which value of second-order roots was higher significantly than that of first-order roots for ash and larch in the two-year period (P＜0.01). By fitting measured each order fine root longevity data to proposed distribution, the appropriate statistical distributions were Welibull, Lognormal, and Normal, seemed as prevenient result, the Exponential distribution still has the inferior fit.(4) The fertilization treatments resulted in an obviously decrease in root number, but a significant increase in root diameter for both total amount and dead amount root in two species (p＜0.01). Nitrogen fertilization treatments also made the root number of explicit order increase in Ash, but Larch had the contrary trend. Contrasting with control treatment, the fit of the Exponential distribution in characterizing fine root longevity was elevated under N amendment treatment, at the same time, the distributions of Welibull, Lognormal, and Normal were still better in character. All above indicated that N fertilization treatment maybe only alter the absolute value of longevity frequency, but the ratio between longevity frequencies was the same as before. So the model of longevity was unalterable.(5) Regardless analyzed with whichever categories (i.e. mortality rate, diameter, season, and soil depth), the median longevity could not exactly estimate the mean longevity except when the mortality rate was low. This result indicated the hypothesis of normally distribution lifespan within fine-root population maybe unadvisable. It would bring biggist error, if we ramstam used median longevity as the mean longevity to estimate fine roots turnover. In unimodal asymmetric distribution, the median lies about one-third the distance between the mean and mode. The equation could be shown as that: mean= (4median- mode)/3. Adopting the experiential equation, we used median and mode to calculate the theoretic mean which was used to estimate the true mean value. By contrasting in different mortality rate, diameter, season, and soil depth categories, we found the theoretic mean could not estimate true mean value exactly as we expected.