Mizue Ohashi

The determination of dead roots is important for clarifying the roles of fine tree roots in carbon (C) and nutrient cycling in forest ecosystems. However, the various methods proposed for determining dead roots and the variation in terminology expressing fine-root mortality can cause significant errors in quantifying dead roots. In this study, we aimed to clarify the various criteria used for sorting dead fine roots using published papers. The second aim was to identify problems in defining fine-root mortality and to clarify the border between defining dead roots and root litter. Finally, we propose future challenges in determining dead fine roots to collect data with the same standards. We collected 95 papers published in the last 50 years and found that visual- and touch-based judgments are most commonly used for sorting dead roots. We found many different terms representing the state of fine-root shedding and mortality. In this study, we proposed that the terminology for dead fine roots needs to be based on the process from root death to decomposition. The accuracy of dead root determination can be improved by experimentally elucidating the fine-root decomposition process. It is also important to increase objectivity in the description of the criteria when judging dead roots in a field. Then, if we combine multiple methods and introducing new tools such as Automatic Intelligence, the mission of identifying dead roots would be possible in future.

To clarify soil respiration (soil CO2 efflux, R-s) characteristics in a subtropical evergreen broad-leaved natural forest in Yambaru, Okinawa, Japan, we examined spatiotemporal variation in R-s and its determining factors. We then compared yearly R-s with the value in other forests. The spatial variation in R-s (coefficient of variation [CV] = 38.9%) was not significantly related to temperature or soil water content but was evidently dependent on ground surface litter coverage. R-s was greater in summer (ca. 7-10 mu mol m(-2) s(-1)), and its seasonal variation was exponentially related to soil temperature (Q(10) = 2.16). As a function of soil temperature, we estimated a yearly mean stand-scale R-s of 5.17 mu mol m(-2) s(-1), and a total carbon efflux from the soil of 1959 g C m(-2) year(-1) for 2014. Despite showing similar seasonal patterns as those in temporal forests, the R-s in this ecosystem is very high throughout the year, and the yearly value is much higher for natural mature forests. A mass balance approach suggests that the large amount of belowground carbon allocation of plants contributed to the high CO2 emissions from the soils.