On the Scaling up Soybean Leaf Level Stomatal Resistance to Canopy Resistance for One-Step Estimation of Actual Evapotranspiration

Abstract

Canopy resistance (rc), which represents the composite diffusive resistance to water vapor transfer from vegetation surfaces to the atmosphere, plays an important role in describing the water vapor and energy fluxes and CO2 exchange mechanisms and is an essential component of the complex ecophysiological and turbulent transport and evapotranspiration models. While one-step (direct) application of combination-based energy balance models (i.e., Penman-Monteith, PM) requires rc to solve for actual evapotranspiration (ETa), a remaining challenge in practical application of PM-type models is the scaling up of leaf-level stomatal resistance (rs) to rc to represent an integrated resistance from the plant community to quantify field-scale evaporative losses. We validated an integrated approach to scale up rs to the canopy. Through an extensive field campaign, we measured diurnal rs for a subsurface drip-irrigated soybean [Glycine max (L.) Merr.] canopy and integrated several microclimatic and in-canopy radiation transfer parameters to scale up rs to rc. Using microclimatic and plant factors such as leaf area index for sunlit and shaded leaves, plant height, solar zenith angle, direct and diffuse radiation, and light extinction coefficient, we scaled up soybean rs as a primary function of measured photosynthetic photon flux density (PPFD). We assumed that PPFD is the primary and independent driver of rc; hence, the scaling approach relied heavily on measured PPFD-rs response curves. We present experimental verifications of scaled up rc by evaluating the performance of the scaled up rc values in estimating ETa. In addition, we solved the PM model on an hourly time step using the scaled up rc values and compared the PM-estimated ETa with the Bowen ratio energy balance system (BREBS)-measured ETa. The relationship between rs and PPFD was asymptotic, and rs showed strong dependence to PPFD, as PPFD alone explained 67% to 88% of the variability in rs. Beyond a certain amount of PPFD (400 to 500 mol m-2 s-1), rs became less responsive to PPFD. At smaller PPFD (0 to about 150 mol m-2 s-1) and greater rs (>70 to 80 s m-1) range, rs was very sensitive to PPFD. The rc_min, rc_avg, and rc_max ranged from 42 to 104 s m-1, 69 to 183 s m-1, and 95 to 261 s m-1, respectively, throughout the season. The seasonal average rc_min, rc_avg, and rc_max were 54, 92, and 129 s m-1, respectively. Canopy resistances were higher in early growing season during partial canopy closure, lower during mid-season, and high again in late season due to leaf aging and senescence. The ETa estimates from the PM model using scaled up rc values correlated very well with the BREBS-measured ETa. The average root mean square difference (RMSD) between the BREBS-measured and PM-estimated ETa was 0.08 mm h-1 (r2 = 0.91; n = 827), and estimates were within 3% of the measured ETa on an hourly basis. On a daily time step, RMSD was 0.64 mm d-1 (r2 = 0.86; n = 83), and the estimates were within 4% of the measured data. The approach successfully synthesized the whole-canopy resistance for use in PM-type combination-energy balance equations by scaling up from rs using a straightforward model of in-canopy radiation transfer.