Seasonal variation of leaf area index (LAI) over paddy rice fields in NE China: Intercomparison of destructive sampling, LAI-2200, digital hemispherical photography (DHP), and AccuPAR methods

•We compared LAI-2200, DHP, AccuPAR, and destructive methods for rice LAI estimation.•Areas of green and yellow leaves, stems, and ears were measured separately.•PAIeff from LAI-2200 and DHP are consistent over the season (R2=0.76, RMSE=0.97).•CI decreases with plant growth and range from 0.63 to 0.74 during the peak season.•PAI and LAI from LAI-2200 and DHP agree with the destructive results (R2>0.75). Continuous field leaf area index (LAI) measurement has become increasingly important for the validation of remote sensing LAI products. A seasonal field campaign was carried out to take continuous LAI measurements over paddy rice fields in NE China in 2012. Three indirect optical methods, LAI-2200, digital hemispherical photography (DHP), and AccuPAR, were compared with a destructive sampling method conducted concurrently. Corrections for the clumping effect were applied to the effective plant area indices (PAIeff) estimated from the indirect optical measurements. Both LAI-2200 and DHP produce consistent PAIeff estimates over the season (R2=0.76, RMSE=0.97). The clumping index (CI) values obtained from DHP generally decrease with plant growth and range between 0.63 and 0.74 during the peak growing period from day of year (DOY) 191−230. The CI values retrieved from DHP photos generally decrease with increasing view angles. The optical PAI and LAI values estimated from LAI-2200 and DHP correspond very well with the destructive values before DOY 230 (R2=0.75, RMSE=1.15 for PAI and R2=0.78, RMSE=0.74 for LAI), and the relative errors are less than 10% and 5%, respectively, for the two instruments. Omitting ring 5 for LAI-2200 generates very accurate PAI and LAI estimations during the peak season. Nevertheless, AccuPAR underestimates the PAIeff, PAI, and LAI values obtained from other methods (up to 30%). After DOY 231, the capability to detect PAI decreases significantly for both destructive and optical methods due to the leaf senescence and the DHP classification difficulty. In general, rice PAI could be accurately estimated with LAI-2200 and DHP before senescence if the clumping effect could be properly taken into account. The seasonal continuous LAI measurements obtained from this study are valuable for the validation of remote sensing LAI products.