Estimating seasonal variations of leaf area index using litterfall collection and optical methods in four mixed evergreen–deciduous forests

Abstract

Leaf area index (LAI), a critical parameter used in process models for estimating vegetation growth, can be measured through litterfall collection, which is usually referred to as a direct method. This method has been demonstrated to be applicable to deciduous forests, but few studies have used this method for estimating seasonal variations of LAI in mixed evergreen–deciduous forests. In this study, we proposed a practical method to estimate the seasonal variation of LAI directly by combining leaf emergent seasonality and litterfall collection (defined as LAIdir) in a mixed broadleaved-Korean pine (Pinus koraiensis) forest (BK), a Korean pine plantation (KP), a spruce–fir valley forest (SV), and a secondary birch (Betula platyphylla) forest (SB). In this direct method, the seasonal variation of LAI in a mixed forest can be quantified by tracking leaf growth and fall patterns throughout the growing season for each major evergreen and deciduous species. Using the LAIdir as a reference, we validated optical LAI (effective LAI, Le) measurements through a digital hemispherical photography (DHP) and the LAI-2000 instrument. We also explored the contribution of major sources of errors to optical LAI, including woody-to-total area ratio (α), clumping index (ΩE), needle-to-shoot area ratio (γE) and automatic exposure (E). We determined that DHP Le significantly (P<0.05) underestimated LAIdir from May to November by 48–64% in BK, KP and SV but overestimated LAIdir by 7% on average in SB. Similarly, LAI-2000 Le also significantly (P<0.05) underestimated LAIdir by an average of 27–35% in BK, KP and SV but overestimated LAIdir by 22% on average in SB. The relative contribution of E to the error in DHP Le is larger than other factors, and the γE was the largest relative contributor to the underestimation of LAI by LAI-2000. The results from our study demonstrate that seasonal variations of LAI in mixed evergreen–deciduous forests can be optically estimated with high accuracy (85% for DHP and 91% for LAI-2000), as long as accurate corrections are made to the various factors mentioned above. These close agreements between direct and optical LAI results also suggest that the direct method developed in this study is useful for tracking the seasonal variation of LAI in mixed forests.