Indirect quantification of fine root production in a near tropical wet mountainous region

The main functions of fine root (defined as diameter <= 2 mm) are water and nutrient transports. Besides being a carbon (C) storage pool, it also provides a C flux pathway through soil and plant. Fine root takes up a small portion, normally 5%, of biomass in forest ecosystems, but 30% to 70% of total net primary production. Therefore, quantifying fine root productivity is important to study the forest C budget. Presumably, belowground growth can be indirectly estimated by the more accessible aboveground vegetation structure dynamics. To verify the relationship with fine root productivity, we take internal (floristic) and external (environmental) factors into account, including litter production, canopy density (leaf area index), leaf nutrients (N, K, Ca, Mg, P), weather and/or soil physical conditions (air temperature, humidity, precipitation, solar radiation and soil moisture). The study was conducted in near tropical broadleaf (700 m asl) and conifer (1700 m asl) forests in northeastern Taiwan, generally receiving more than 4000 mm of precipitation per year. For each site, 16 50-cm long minirhizotron tubes were installed. Fine root images were acquired every three weeks. Growth and decline, newly presence and absence of fine roots were delineated by image processing algorithms to derive fine-root productivity through time. Aforementioned internal and external attributes were simultaneously collected as well. Some of these variables were highly correlated and were detrended using principal component analysis. We found that these transformed variables (mainly associated with litter production, precipitation and solar radiation) can delineate the spatiotemporal dynamics of root production well (r2 = 0.87, p = 0.443). In conclusion, this study demonstrated the feasibility of utilized aboveground variables to indirectly assess fine root growth, which could be further developed for the regional scale mapping with aid of remote sensing.