On the dynamics of stomatal resistance: Relationships between stomatal behavior and micrometeorological variables and performance of jarvis-type parameterization

Two Jarvis-type models, the original Jarvis model (J-model) and the Green and McNaughton model (GM-model), were evaluated for estimating leaf stomatal resistance (rs) for a non-stressed maize canopy. The parameters of the models were optimized, and their performances were compared against measured rs. A new modified Jarvis-type model (NMJ-model) was developed by including a new term, r _s-min exp(-LAI), where r _s-min = minimum stomatal resistance (sm ¹) and LAI = green leaf area index, to account for the influence of canopy development stage on rs, especially during partial canopy. Through an extensive field campaign, hourly rs, photosynthetic photon flux density (PPFD), air and leaf temperatures (Ta and TL), porometer cup temperature (Tc), air relative humidity above the canopy (RHa) and on the leaf surface (RHL), wind speed (u3) and direction at 3 m, vapor pressure deficit (VPD), LAI, and incoming shortwave radiation (Rs) were measured for a subsurface drip-irrigated non stressed maize canopy. The relationship between rs and Ta, TL, Tc, RHa, VPD, Rs, u3, and wind direction are presented. On a seasonal average basis, the J-model and the NMJ-model had similar performance in estimating rs (r ² = 0.74, RMSD = 48.8 sm ^-1 for the J-model; and r ² = 0.74, RMSD = 50.1 sm ^-1 for the NMJ-model; RMSD = root mean square difference between modeled and measured Rs). The inclusion of the variation in the LAI and r _s-min term (r _s-min exp(-LAI)) during the growing season in the NMJ-model improved the rs estimates, especially in the higher rs range (rs > 250 s m ^-1), as compared with the J-model. When the period for only partial canopy cover is considered, when LAI ranged from 1.20 to approximately 2.5, the addition of the LAI term in the NMJ-model resulted in 8% improvement in r ² and 10% improvement in RMSD relative to the J-model (r ² = 0.64, RMSD = 35.5 s m ^-1 for the NMJ model; and r ² = 0.59, RMSD = 39.0 sm ^-1 for the J-model) when estimating rs. The estimated rs using T _L rather than Ta correlated very well with the measured values, but with a slight decrease in performance for both the J model and the NMJ-model. Overall, the performance of the J-model and the NMJ-model decreased by 17% when T _L rather than T _a was used to estimate rs as compared with the measured rs. While the results demonstrated the superior performance of the NMJ-model over the J-model, especially during the partial canopy conditions, its performance needs further investigation and validation in other climatic settings.