Houborg_2015a

Leaf chlorophyll constraint on model simulated gross primary productivity in agricultural systems

​Houborg R, McCabe MF, Cescatti A and Gitelson AA (2015)
International Journal of Applied Earth Observation and Geoinformation, 43, 160-176
​Houborg R, McCabe MF, Cescatti A, Gitelson AA
Landsat; Leaf chlorophyll content; Vmax; Leaf photosynthetic capacity; Community Land Model; Agriculture; Rubisco; Nitrogen
2015
Leaf chlorophyll content (Chll) may serve as an observational proxy for the maximum rate of carboxylation (Vmax), which describes leaf photosynthetic capacity and represents the single most important control on modeled leaf photosynthesis within most Terrestrial Biosphere Models (TBMs). The parameterization of Vmax is associated with great uncertainty as it can vary significantly between plants and in response to changes in leaf nitrogen (N) availability, plant phenology and environmental conditions. Houborg et al. (2013) outlined a semi-mechanistic relationship between View the MathML source(Vmax normalized to 25 °C) and Chll based on inter-linkages between View the MathML source, Rubisco enzyme kinetics, N and Chll. Here, these relationships are parameterized for a wider range of important agricultural crops and embedded within the leaf photosynthesis-conductance scheme of the Community Land Model (CLM), bypassing the questionable use of temporally invariant and broadly defined plant functional type (PFT) specific View the MathML source values. In this study, the new Chll constrained version of CLM is refined with an updated parameterization scheme for specific application to soybean and maize.

The benefit of using in-situ measured and satellite retrieved Chll for constraining model simulations of Gross Primary Productivity (GPP) is evaluated over fields in central Nebraska, U.S.A between 2001 and 2005. Landsat-based Chll time-series records derived from the Regularized Canopy Reflectance model (REGFLEC) are used as forcing to the CLM. Validation of simulated GPP against 15 site-years of flux tower observations demonstrate the utility of Chll as a model constraint, with the coefficient of efficiency increasing from 0.91 to 0.94 and from 0.87 to 0.91 for maize and soybean, respectively. Model performances particularly improve during the late reproductive and senescence stage, where the largest temporal variations in Chll (averaging 35–55 μg cm−2 for maize and 20–35 μg cm−2 for soybean) are observed. While prolonged periods of vegetation stress did not occur over the studied fields, given the usefulness of Chll as an indicator of plant health, enhanced GPP predictabilities should be expected in fields exposed to longer periods of moisture and nutrient stress. While the results support the use of Chll as an observational proxy for View the MathML source, future work needs to be directed towards improving the Chll retrieval accuracy from space observations and developing consistent and physically realistic modeling schemes that can be parameterized with acceptable accuracy over spatial and temporal domains.