Spatial And Temporal Distribution Of Photosynthetic Active Radiation Within Cotton Canopy Using Digital Images

The objectives of this study were to investigate the spatial-temporal heterogeneity of PAR indifferent plant densities during growth stage and to establish PAR models based on color features. Sixdensity treatments were arranged to generate a range of different canopy structure and light environment.Cotton growth and development characteristics and environment factors were considered, and spatialstatistics and digital image technology were applied to study the spatial pattern and spatial-temproalheterogenity. Based on the trend of color features over growth period, correlation between LAI,PARtransmittance rate and color features were analyzed, and PAR transmittance models using colorfeatures were eatablished.The results showed that PAR transmittance rate varied moderately in different growth stages andthe variation degree increased with increasing plant density, however, the means decreased withincreasing density. The PAR transmittance rates of all treatments perfectly fitted the Gaussiansemivariogram model. The structural analysis indicated that PAR transmittance rate had good spatialcontinuity and strong spatial autocorrelation, which mainly caused by structure factors. Sill valueincreased first and then experienced a stable phase and finally declined. C/(C+C0) increased before fullbud time, and then tend to be a constant,indicating that PAR transmittance rate had spatial and temporalheterogenity.The contour map of PAR in different growth stages showed "V" or "U" distrbution pattern. Lightintensity within the canopy was weakening with increasing accumulated LAI, but the weakening ratesin central and top layers were greater than that in the bottom layer and the weakening rates had agrowing tendency with increasing plant density. For greater densities (8.7×104å'Œ1.05×105plants hm-2),there was an obvious difference of congested degree in contour line between top and bottom layerwithin canopy, showing stronger vertical spatial variation. In contrast, the spatial variations were theweakeast in lower densities (1.5×104å'Œ3.3×104plants hm-2) and weaker in middle densities (6.9×104å'Œ8.7×104plants hm-2). The variation of PAR in same sampling height was stronger in bud stage (Prior tocanopy closure) than in blooming-bolling stage.LAI, plant height, dry matter and POW/DW were used as main canopy structure indices. Theheterogenity of canopy structure was increasing as structure indices increased in bud stage, and theheterogenity degree increased with increasing density. PAR transmittance rate declined and PARcaptured was less and PAR diurnal tended to be a constant. Entering blooming-bolling stage, LAI had acontinuous increase until peak full bolling stage and the canopy structure index still increased whichcaused the development of canopy structure and reached the greatest heterogentiy degree to canopy.PAR interception amount was great and PAR transmittance rate was continuing decline. Fullbloomstage, the rage of PAR diurnal variation was increased, and PAR transmittance rate was great in lowerdensity than in middle and higer densities.Full boll stage, PAR transmittance rate showed the annualchange. Boll opening stage, heterogenity of canopy declined resulting from the reduced LAI, which caused PAR transmittance rate slightly increased and PAR diurnal variation was still great. Thegreatest value of PAR diurnal variation was appeared around12:00.There was significant correlation between color feature and LAI,PAR transmittance rate. H and H2showed the greatest correlation, and the LAI and PAR model using these two color features, whichdegree of fitting was high and RMSE was less and can be used to describe PAR transmittance rate indifferent stage acurately.