Uncertainties in fPAR estimation of grass canopies under different stress situations and differences in architecture

Abstract

The fraction of intercepted photosynthetic active radiation (fPAR) is a key variable used by the Monteith model to estimate the net primary productivity (NPP). This variable can be assessed by vegetation indices (VIs) derived from spectral remote sensing data but several factors usually affect their relationship. The objectives of this work were to analyse the fPAR dynamics and to describe the relationships between fPAR and several indices (normalized difference vegetation index (NDVI), optimized soil adjusted vegetation index (OSAVI), Green NDVI (GNDVI), visible atmospherically resistant index (VARI) green, VIgreen and red edge position (REP)) under different water and nutrient treatments for two species with different canopy architectures. Two C 3 grass species with differences in leaf orientation (planophile and erectophile) were cultivated from seeds in pots. Four treatments were applied combining water and nitrogen availability. Every week, canopy reflectance and fPAR were measured. Aerial biomass was clipped to estimate final above-ground production for each species and treatment. Starting from reflectance values, the indices were calculated. Planophile species have a steeper (but not significantly) slope in VIs-fPAR relationships than the erectophile species. Water and nutrient deficiencies treatment showed no relationship with fPAR in any spectral index in the erectophile species. In the other species, this treatment showed significant relationship according to the index used. Analysing each species individually, treatments did not modify slopes except in one case (planophile species between both treatments with high nitrogen but differing in water availability). Among indices, GNDVI was the best estimator of fPAR for both species, followed by NDVI and OSAVI. Inaccurate results may be obtained from commonly reported spectral relationships if plants' stress factors are not taken into account.

Modified vegetation indices for estimating crop fraction of absorbed photosynthetically active radiation

“…However, the vegetation species have significant influence on most VIs and hence limit their applications (Myneni and Williams 1994;Viña et al 2011). Regression results in Table 3 and Table 4 showed that FPAR of different crops were correlated differently with VIs, consistent with the studies by Cristiano et al (2010) and Viña and Gitelson (2005). This suggests that under the same atmospheric conditions different crop types with different canopy architectures might have the same FPAR but different LAI.…”

“…Different VIs have been used for estimation of FPAR, and an ideal VI should be sensitive to FPAR through the whole dynamic range. However, the relationships between most VIs and FPAR are not stable due to impacts from several external factors such as the illumination geometry (Epiphanio and Huete 1995;Ogutu and Dash 2013;Roujean and Breon 1995), canopy background reflectance (Goward and Huemmrich 1992;Huemmrich and Goward 1997), and canopy structure parameters (Asner and Wessman 1997;Cristiano et al 2010;Dong et al 2015;Goel and Qin 1994;Goward and Huemmrich 1992). The relationships are also significantly affected by the saturation effect (Di Bellat et al 2004;Ridao, Conde, and Minguez 1998;Viña and Gitelson 2005).…”

“…The green NDVI (GNDVI) (Pinter 1993), with the red band reflectance replaced by the green reflectance, exhibits an increased sensitivity to moderate-to-high FPAR in comparison to NDVI; however, its sensitivity decreases as FPAR continues to increase (FPAR > 0.8) (Gitelson, Kaufman, and Merzlyak 1996;Viña and Gitelson 2005). The study by Cristiano et al (2010) showed that the red-edge NDVI (NDVI red-edge ) and the wide dynamic range VI (WDRVI) (Viña and Gitelson 2005) have the highest sensitivity to green FPAR over the entire variation range in a growing season. The red-edge reflectance is more sensitive to high chlorophyll content than the red band reflectance (Gitelson, Kaufman, and Merzlyak 1996).…”

“…This leads to saturation in VIs that rely on the contrast between the red and the NIR reflectance, and hence degrades the sensitivity to FPAR at moderate to high biomass (Gitelson, Kaufman, and Merzlyak 1996;Ridao, Conde, and Minguez 1998;Viña and Gitelson 2005). Much effort has been focused on modifying the existing VIs or developing new VIs to reduce the 3098 T. Dong et al uncertainties in biophysical parameter estimation (Cristiano et al 2010;Roujean and Breon 1995;Viña and Gitelson 2005). For example, the renormalized difference VI (RDVI), which combines the advantages of the difference VI (DVI) and NDVI for low and high vegetation cover fraction, respectively, has proved to be better for FPAR estimation, because it is more sensitive to low vegetation cover and less affected by the effect of solar angle than the other VIs (Roujean and Breon 1995;Zhang et al 2009).…”

Assessment of MODIS, MERIS, GEOV1 FPAR Products over Northern China with Ground Measured Data and by Analyzing Residential Effect in Mixed Pixel

“…It is very important to collect reliable ground measured FPAR data for product validation and even for its utilization in modeling communities [46]. At present, a few instruments have been widely used, such as the Decagon AccuPAR ceptometer [15,23], tracing architecture and radiation of canopies (TRAC) [20], LI-191SA line quantum sensors or LI-190SZ (LI-COR Inc., Lincoln, NE, USA) [47,48], TM Cava devices line quantum sensor [49] and SKYE PAR quantum sensors (SKP 215 and SKR 110) [21]. Serbin et al [19] found MODIS C5 FPAR product follows well the temporal evolution of LAI-191SA FPAR data.…”