Variations in the leaf area index and its effect on estimations of primary production in a natural hardwood forest and a Cunninghamia lanceolata plantation at the lienhuachi experimental forest, central Taiwan

We set up 3 crest-to-valley transects each in watersheds no. 3 (a natural hardwood forest) and no. 4 (a Cunninghamia lanceolata plantation) at Lienhuachi Experimental Forest of central Taiwan in 2005. We measured the leaf area index (LAI) 6 times using a LI-COR LAI-2000 Plant Canopy Analyzer. The mean LAI measured during the sampled year ranged from 2.64 to 4.72 in the natural hardwood forest and from 3.18 to 3.95 in the C. lanceolata plantation. Both types of forests showed substantial intra-annual variation. Unusually low temperatures in March caused a 31% decrease in the LAI in the hardwood forest, but in the C. lanceolata plantation, the decrease was not detected until May and it was much smaller (15%). After typhoon Haitang passed nearby on July 18-19 LAI values decreased 11 and 18% in the hardwood forest and C. lanceolata plantation, respectively. Typhoon Haitang seemed to have a greater impact on the C. lanceolata plantation than on the natural hardwood forest. For the 6 samplings, the largest differences in LAI among the 3 transects reached 3.16 (November) in the hardwood forest in which the LAI of B3 (6.34) was approximately 2 times that of B2 (3.18). For the C. lanceolata plantation the largest difference reached 3.12 (March) with an LAI of C3 (5.73) being more than 2 times that of Cl (2.62). LAI differences among transects were greater in the C. lanceolata plantation than in the natural hardwood forest. Applying the measured LAI values to an ecological model showed that the temporal variation in LAI led to 66 and 22% differences in the estimation of the above-ground net primary production (ANPP) of the natural hardwood forest and C. lanceolata plantation, respectively, and spatial variations in LAI led to 45 and 49% differences in ANPP estimations. Without considering such temporal and spatial variations in LAI, estimations of ANPP will be dependent on both time and location, and such variations can be greater than differences among forests of different types and at different sites.