Influence of automatic exposure on the estimation of seasonal variations in leaf area index measured by digital hemispherical photograph.

Automatic exposure is one of the important error sources during measurement of leaf area index (LAI) by digital hemispherical photography (DHP). This study was conducted in a mixed broadleaved-Korean pine (Pinus koraiensis) forest, a secondary birch (Betula platyphylla) forest, a Korean pine plantation and a Dahurian larch (Larix gmelinii) plantation in the Xiaoxing'an Mountains. LAI was measured using DHP and LAI-2200 plant canopy analyzer in the middle of June to September. We compared LAI values measured through these two methods, and then tested whether the forest type and study period had a significant influence on the correlations between the measured values of those two methods. We constructed empirical models for correcting the errors caused by automatic exposure for LAI values measured through DHP at different study periods in different forest types. The results showed that LAI from DHP was underestimated by 20%-49% rela-tive to that from LAI-2200 in four study periods of the four forest types. Forest type had no significant effect on the construction of empirical models between these two measuring methods of LAI, whereas study period showed significant effects. Two classified empirical models (A and B) were constructed, which were suitable for correcting the LAI from DHP in June and September, July and August in four forest types, respectively. After being corrected by the classified empirical models, LAI from DHP of the four forest types increased by 45%-79%, and the measurement accuracy could be improved to 83%-94%. Classified empirical models between LAI from DHP and LAI-2200 could effectively correct the influence of automatic exposure on DHP and greatly improve its mea-surement accuracy, and provide a technical support for rapid and effective measurement of seasonal changes of LAI in different forest types.