1. Introduction
Land surface processes (LSPs) and their interactions with the atmosphere are critical at different scales (Pielke et al. 2002; Feddema et al. 2005; Holt et al. 2006; Pielke and Niyogi 2009). Hence, LSPs need to be modeled as a continuum involving coupled feedback between different biospheric and hydrometeorological components. LSPs are modeled through soil–vegetation–atmosphere transfer (SVAT) schemes in which the stomatal response is the key to developing realistic biosphere–atmosphere interactions [see Niyogi and Raman (1997) for a review].
In the 1970s, most LSP parameterizations were diagnostic with an emphasis on the bare ground processes (e.g., Blackadar 1976). The soil–vegetation scheme proposed by Jarvis (1976) and Deardorff (1978) represents the “second generation” of the SVAT models. The Jarvis-type evapotranspirative schemes are relatively simple and dependent on meteorological variables such as air temperature and incident radiation. With several empirical modifications, the Jarvis schemes (Jarvis 1976) continue to be adopted to estimate latent heat fluxes and the surface energy balance in atmospheric models (Noilhan and Planton 1989; Chen et al. 1996; Chen and Dudhia 2001; Ek et al. 2003). However, as discussed in Niyogi et al. (1998), these second-generation SVAT schemes do not rigorously account for the coupled vegetation–atmosphere interactions and boundary layer feedback processes important to various environmental applications (Niyogi and Raman 2001). These limitations are addressed in the CO2-based physiological or “third generation” land surface models (LSM; Sellers et al. 1996; Pitman 2003) that consider evapotranspiration an inevitable cost during photosynthesis (Cowan 1982).
The third-generation SVATs are relatively complex schemes that consider photosynthesis (carbon assimilation) and biochemical responses and are explicitly coupled with the atmosphere and the surface to estimate carbon, water vapor, and energy exchange (Farquhar et al. 1980; Collatz et al. 1991, 1992; Anderson et al. 2000; see also Niyogi and Raman 1997). These schemes generally have been applied in leaf/canopy scale models (e.g., Baldocchi 1992), or in global/regional climate studies (Sellers et al. 1996; Calvet et al. 1998; Cox et al. 1999; Dai et al. 2003). Mesoscale or weather forecast models continue to be predominantly Jarvis-based (e.g., Chen and Dudhia 2001; Ek et al. 2003), but several reasons support the adoption of the photosyn thesis schemes over the Jarvis-type approaches.
First, for the estimate of stomatal resistance (Rs), the Jarvis approach relies on the prescription of so-called minimum stomatal resistance (Rs,min). However, the Rs,min prescription not only varies with vegetation type, but also varies with the seasons (Schulze et al. 1994; Alfieri et al. 2008). Second, as shown in Alapaty et al. (1997a) and Cooter and Schwede (2000), Rs,min errors impact surface hydrology and boundary layer simulations. These photosynthesis-based relations, though not causal, are based on gas exchange parameters that can be measured in the laboratory or in field studies (Ball et al. 1987). The gas exchange parameters seem to be robust and show phenological similarities without any significant seasonal or other environmental modulation in the basal values (Collatz et al. 1991). Third, with the current advances in remote sensing technology, a number of high-resolution land use–land cover and vegetation phenology datasets now can facilitate applications of photosynthesis-based models (Masson et al. 2003). Additionally, the photosynthesis-based schemes are more interactive and can thus better replicate the stomatal resistance variations under a variety of environmental changes (Niyogi et al. 1998).
In this study, we show first that the photosynthesis-based gas exchange transpiration model (GEM), despite various complexities, can be efficiently coupled to boundary layer and mesoscale models, and is preferable over the Jarvis scheme. In section 2, we describe the photosynthesis- based gas exchange evapotranspiration model (GEM) formulations. In section 3 we describe the coupling of the GEM with a prognostic soil moisture/soil temperature scheme and an atmospheric/PBL model; we also describe the development of initial conditions for the different model evaluation case studies. Results from the coupled model, for nine different cases, are then discussed in section 4. The conclusions from this study are summarized in section 5.
2. Model description
The photosynthesis-based gas exchange evapotranspiration model that we developed was integrated within a prognostic soil moisture and soil temperature prediction model similar to the one developed by Noilhan and Planton (1989) and Alapaty et al. (1997a). A brief summary of the model components is presented below. The equations and the parameterization schemes are described in the appendixes.
In the GEM, the estimation of the different parameters required to solve the Ball–Berry model [Eq. (1)], particularly the carbon assimilation (An) term, mainly follows the Collatz et al. (1991, 1992) scheme. Also, the coupling between the hs and Cs is achieved in solving for the photosynthesis rate estimation, following Sellers et al. (1996). The stomatal resistance calculations are interactively linked with soil variables such as soil temperature and soil moisture as well as with the atmospheric components such as wind, radiation, pressure, temperature, and humidity. The photosynthesis-based stomatal resistance model is linked to a Noilhan and Planton (1989) type prognostic soil moisture–soil temperature (SMST) scheme with two soil depths (0.1 and 1 m) and one canopy layer. Noting that the model will be typically applied at mesoscale model grids with ∼10-km horizontal spacing, a number of simplifying assumptions have been made in the interest of computational efficiency. These include using surface temperature from the SMST scheme (which accounts for the fractional vegetation cover) in place of an explicit canopy temperature calculation as well as the consideration of big-leaf scaling instead of the multilayer canopy model. Recently, the GEM has been coupled within the Noah land surface model with four soil layers, and similar formulations can be used (cf. Kumar et al. 2007, 2008). The SMST scheme is coupled as a surface energy balance scheme to an atmospheric boundary layer model as described in Alapaty et al. (1997a). To couple the models, the GEM Rs output is linked with the surface energy balance model that is coupled to the atmospheric boundary layer model. In other words, the Jarvis-type stomatal resistance is replaced with the GEM-based estimates. All three schemes, the GEM, the soil model, and the PBL model, are interactive at every time step (∼30 s).
3. Data and numerical simulations
The coupled SVAT scheme discussed above is primarily developed for land surface models that can be coupled to mesoscale numerical weather forecast models. Since the scheme is being adopted in a 3D modeling framework such as the Noah Model within the Weather Research and Forecasting (WRF) Model (e.g., Kumar et al. 2008), a number of tests were necessary to evaluate the model’s ability to reproduce important features related to surface boundary conditions, such as surface energy fluxes. In this study, we provide an overview of the model development and report on tests over different land covers by comparing the modeled and observed surface energy fluxes under contrasting micrometeorological conditions. These land covers were selected based on the following criteria: (i) test of two different photosynthesis submodels for the C3 and C4 pathways, (ii) tests over grasslands, agricultural crops, and forests (the most common vegetation types), (iii) availability of soil moisture and plant biophysical data, and (iv) our experience working with the site scientists and the datasets, which reduced uncertainties in defining the model initial conditions.
The study sites included (i) a temperate prairie grass land with a mixture of C4 vegetation, (ii) a C4 corn crop field, (iii) a C3 soybean crop field, (iv) a mixed C3 natural fallow field, and (v) a mixed C3 forest area. For all the sites, except the fallow field (for which only one day of data was readily available), two case studies were performed under different soil moisture and/ or vegetation characteristic conditions, a total of nine scenarios. The soil–vegetation model also requires the initial condition to be externally specified. The values of the surface variables used to simulate this case are shown in Table 1. We recognize that the single-day coupled simulations provide only an initial analysis of the model performance. A longer simulation in an offline mode is currently under way by coupling the GEM formulation within the Noah land surface model as part of the National Center for Atmospheric Research (NCAR) high-resolution land data assimilation system, and the initial results are consistent with those reported here (cf. Kumar et al. 2007, 2008).
For the prairie grassland case [First International Satellite Land Surface Climatology Project (ISLSCP) Field Experiment (FIFE1 and FIFE2 data)], Alapaty et al. (1997a,b) developed a thermodynamically consistent, calibrated SVAT input for a Jarvis-based scheme following Noilhan and Planton (1989). The corresponding model results were validated using previously published results and field observations; the calibrated results were within 10%–20% of the observations (which in themselves had equivalent scatter). Hence, we will use the model output for this case as a surrogate for variables such as stomatal resistance, which are not directly observed for most field studies. The different cases are discussed in the following sections.
We defined the FIFE (Sellers et al. 1992) site as dense, temperate C4 grassland in Kansas. The FIFE site covers a region of 15 km × 15 km near Manhattan, Kansas (Polley et al. 1992; Niyogi and Raman 1997). We discuss the case for intensive field campaigns (IFCs) as the “golden days 1 and 2”: 7 June and 11 July 1987. Various researchers have extensively tested the golden days, and a reference input dataset is available [see, e.g., Sellers et al. (1992), and a series of papers included in that issue, as well as Alapaty et al. (1997a,b), Bosilovich and Sun (1995), and Niyogi and Raman (1997)]. Numerical simulations were performed for a period of 12 h starting at 1200 UTC (0700 LT) for both cases.
The second evaluation case was performed for a homogeneous C4 crop canopy (corn). Two case studies were selected: CORN-FULL (a fully grown stage with relatively low soil moisture) and CORN-OLD (the drying crop stage with relatively higher soil moisture). Following the two evaluations over C4 vegetation, additional case studies were undertaken over C3 vegetation. The first two of these were over a soybean canopy in an agricultural field: SOY-FULL (a fully grown soybean crop) and SOY-OLD (older, drying soybean crop). Of these two case studies, the soil moisture availability was high for the SOY-FULL case, and was low for the SOY-OLD case. Details regarding these field observations can be found in Meyers et al. (1998).
One of the simplified aspects of the GEM, as a photosynthesis-based SVAT scheme for atmospheric models, is the traditional big-leaf scaling from the leaf to the canopy on the basis of leaf area index (LAI). Therefore, two case studies were performed over a forested site. Finkelstein et al. (2000) made micrometeorological observations of surface energy fluxes, along with other environmental variables at a forest site in Sandflats, New York. For this forested site, two cases were simulated: FOREST-WET (high soil moisture period) and FOREST-DRY (low soil moisture period). The corresponding input surface data are shown in Table 1.
In the case studies, the evaluation parameters are the surface heat fluxes. Evaluating fluxes, instead of individual evapotranspiration components, is sufficient and appropriate considering GEM is intended to be used as a surface energy balance scheme in the land surface models. The sensible (SHF) and latent heat fluxes (LHF) were measured using an eddy covariance technique and were quality assured by site investigators. In addition to the fluxes, we also compared the model results with the best estimates of stomatal resistance (Rs) for the FIFE case, and the observations of photosynthesis rates (An) for the Monitoring the Usable Soil Reservoir Experimentally (MUREX) case (Calvet et al. 1999). These two parameters, Rs and An, are key variables in GEM.
For all of the cases, with the exception of the FIFE, on-site sounding data were not available. In the 1D model, meteorological sounding data were used to initialize the ABL model with representative wind and thermodynamic structure and to provide external boundary conditions for winds. The observed sounding data were also useful for evaluating the model-simulated boundary layer variables. Hence, when direct on-site observations were not available, archived analysis-based upper-air data were obtained corresponding to the study’s time and location. These sounding products were obtained through the National Oceanic and Atmospheric Administration (NOAA) Realtime Environmental Analysis and Display System site (http://www.arl.noaa.gov/ready.html) and the University of Wyoming (http://www.das.uwyo.edu). The sounding data were then processed for geostrophic winds and interpolated to the model sigma levels. These vertically integrated soundings were then temporally interpolated to generate hourly data over the numerical simulation locations (Alapaty et al. 2001). These processed data were used for initialization and developing boundary conditions for the atmospheric model.
4. Results and discussion
The simulation results over the different landscapes are presented in this section. The results are presented first for the C4 vegetation and then for the C3 vegetation.
a. Simulation over C4 vegetation
In this section, the model results from the FIFE cases over temperate grassland and over a corn field are discussed. Both the vegetation canopies showed a C4 photosynthesis pathway.
1) FIFE1
This case corresponded to the FIFE golden day 1 (Sellers et al. 1992) on 6 June 1987. During this period, the vegetation was in its “greening up” phase with moderate to low soil moisture availability. Figures 1a–c show the observed and the predicted net radiation (Rn), latent heat fluxes, and sensible heat flux for this case. Though the modeled net radiation flux estimates were not directly linked with the transpiration model, the agreement of these estimates with the observations provided confidence in the specified initial conditions of the biophysical parameters such as albedo, emissivity, and soil texture (Table 1). The simulations closely followed the observed Rn, generally within 20 W m−2, until early evening. After 1600 LT the observations were larger than the model-estimated values by about 30–40 W m−2. The model-predicted latent and sensible surface heat fluxes are also in good agreement with observations. The latent heat fluxes, in particular, are generally within about 20 W m−2 of the observations. For most of the day, the model-predicted values were slightly lower than the observations and the afternoon and early evening predictions were about 25 W m−2 higher. As expected for the vegetated field with freely transpiring surface, the SHF values are lower than the LHF. The SHF values show some scatter in the morning, and reach a peak of about 130 W m−2. At the beginning of the day, the simulated values were within about 20 W m−2 until the afternoon, and then they were within about 35 W m−2 into the later part of the day. The modeled SHF values are underestimated in the morning and evening and overpredicted in the afternoon. This could be due to vegetation stress in the afternoon leading to lower-than-observed LHF (in the latter part of the day) and uncertainty in the input conditions. Overall, the coupled scheme simulated the surface energy fluxes fairly well, with values within 10% of the observations.
In addition to the surface fluxes, observed soundings are available for FIFE. We evaluated the model-simulated vertical virtual potential temperature profiles against the observations. As shown by Alapaty et al. (1997a), surface fluxes can impact the entire boundary layer structure, which corroborates the importance of this study of GEM coupled to an atmospheric boundary layer model. The vertical distribution of the virtual potential temperature (θυ) and the resulting PBL height variation provides an overall view of the mean thermodynamic structure of the boundary layer. Figure 2 shows a summary of the observed and model-predicted boundary layer heights. The observed boundary layer deepened over the day with the 0700, 1000, 1300, and 1600 LT values corresponding to about 250, 1000, 1300, and 1400 m, respectively. The evolution of the observed and the model-simulated θυ profiles (not shown) also shows good agreement with each other. At 0700 LT, the observations indicated a shallow (∼250 m) stable boundary layer. This profile was used as the model initial condition. After model integration, around 1000 LT, the PBL was weakly unstable with mean temperatures in the mixed layer around 299 K. At 1300 LT, the boundary layer evolved further with a mixed layer θυ value around 302 K. The boundary layer was warmer in the 1600 LT profile with a surface value of 305 K. The simulated θυ values were similar to the observations, though lower by about 1 K in the PBL. The θυ values and the boundary layer depth are often closely linked with the surface heat fluxes, since the atmospheric eddy size increases with higher sensible heating, thereby deepening the boundary layer. If more validation data had been available, the relationship between the errors in θυ and errors in the SHF could have been verified. In general, the model simulations are fairly similar to the observations.
2) FIFE2
The FIFE2 case corresponds to the FIFE golden day 2 on 11 July 1987 (Sellers et al. 1992). As compared with FIFE1, FIFE2 corresponds to modestly higher initial soil moisture availability (0.27 as compared with 0.25 m3 m−3 in the top 10-cm soil layer). The vegetation is also at its peak green with a higher LAI than in FIFE1 (2.8 as compared with 1.9). Similar to the FIFE1 case, this day corresponds to clear-sky conditions with nearly advection-free meteorological conditions for the initial part of the day. However, after 1500 LT, overcast-sky conditions were recorded; in the absence of explicit large-scale advection, the 1D model results are ideally comparable with observations only until 1500 LT.
Figures 3a–c show the temporal variations in the observed and simulated net radiation and surface sensible and latent heat fluxes. The simulated SHF and LHF values are also within 10%–20% of the observations and are within the range of uncertainty in the model formulations and the measurements (LeMone et al. 2007). The model appears to reproduce the observations in terms of the temporal evolution as well as the measured quantities for the temperate prairie grassland. Also, for this case, the model performance is relatively better than FIFE1. This could be due to a number of factors including a denser and more homogeneous vegetation canopy, as well as modestly higher soil moisture availability (see also Colello et al. 1998). This aspect of better model performance under higher soil moisture availability is seen as a general conclusion for the different cases discussed ahead.
The model-estimated and observed boundary layer heights are plotted in Fig. 4. The observed and model-predicted θυ profiles also agree well (not shown). The model evolution of the PBL height and potential temperature are consistent with the good performance of the ABL model over the FIFE region reported by Alapaty et al. (1997a). Thus the coupling of GEM within the LSM appears to be dynamically consistent within the coupled SVAT model that was initially tested and calibrated using a Jarvis-type vegetation representation by Alapaty et al. (1997a) and Niyogi et al. (1999).
3) Corn-high
This case corresponds to a period when the corn crop was at its peak with an LAI around 3 (Meyers et al. 1998). The soil moisture availability was moderate to low. The model was initialized at 0200 LT 24 August 1994 and integrated for a period of 48 h. The observed and the model-simulated net radiation and the sensible and latent heat fluxes are shown in Figs. 5a–c. The model errors are on average within 20% of the observations, with the spikes showing larger differences. It is interesting to note that the model is able to simulate the day-to-day variations in the energy fluxes noted in the observations. The model shows much more negative net radiation at night than seen in the observations. This nighttime net radiation error is a known problem with the SVAT scheme used in Alapaty et al. (2001). Sensitivity runs (not shown) suggest that the error may be mainly due to the prescribed albedo and emissivity values. In the current version, both albedo and emissivity are independent of soil moisture and are one of the modifications we plan to incorporate in a follow-up study for integration within the Noah model (Idso et al. 1975; Ek et al. 2003; LeMone et al. 2008). Also, we do not know the possible impact of correcting the observations to account for the energy balance closure (Wilson et al. 2002).
4) Corn-old
This case study for 15–16 September 1994 simulated the atmosphere over the same crop field simulated in the CORN-FULL case. In this case, the corn was aged and drying primarily because of senescence. The LAI was about half of the CORN-FULL case, but the soil moisture availability was much higher. This created a confounding between the wilting vegetation that generally resists transpiration with age (Niyogi et al. 1997) and the high–soil moisture availability that induces higher–water vapor exchange through the stoma. Figures 6a–c show the temporal variations in the observed and the model-predicted Rn, SHF, and LHF. Largely due to scattered (subgrid) clouds over the study domain, the observed net radiation values show significant variability particularly in the late afternoon and early evenings for both days. The error in Rn introduces some inherent limitations in simulating SHF and LHF values. Generally, the effect of the Rn overprediction will be more on the SHF than the LHF model predictions. Although the LHF values depend on Rn, the error results because these values were calculated explicitly through the GEM-based SVAT scheme, and the SHF prediction is a function of the air and land surface temperature differential. Thus, the simulated LHF was lower than the observations by about 10%, while the sensible heat fluxes were generally overestimated by about 20%. Despite these differences and considering the difficulties of simulating this case with limited site-specific input data, the overall coupled model performance was quite satisfactory for the C4 vegetation. Interestingly, even though the vegetation was wilting, the overall model performance appears to be comparable with the CORN-FULL case, possibly due to the high soil moisture availability.
b. Simulation over C3 vegetation
This section discusses the numerical simulations and evaluation results over three different C3 vegetation types: a soybean crop, a fallow field, and a hardwood forest.
1) Soy-full
During this case, the soybean crop was at its peak green with LAI around 6 and high soil moisture availability (Meyers et al. 1998). The model was initialized at 0600 LT 6 August 1995 and integrated for a 48-h period with the surface conditions shown in Table 1. Figures 7a–c show the observed and the modeled Rn, SHF, and LHF for the 2 days, which are similar to the results seen for the CORN case and therefore are not further discussed here in detail.
2) Soy-old
This case study was performed over the same soybean crop as discussed above, but for a period about 5 weeks later than the peak green, and the vegetation was aged and partly wilting (Meyers et al. 1998). The model simulations involved a 48-h period starting from 0600 LT 12 September 1995. Figures 8a–c show the observed and the modeled radiation fluxes. The Rn observations showed significant variability due to intermittent cloud cover. The afternoon peak for the 2 days was around 350 W m−2 for Rn, which the model captured well. However, the observed and modeled Rn values differed by about 50 W m−2 for any individual hour. The maximum LHF and SHF values measured approximately 150 and 100 W m−2, respectively. Though the simulations did not show the variability seen in the open canopy, the gross features including day-to-day variability were captured. Results also indicate that short-term errors can occur for such a modeling case when vegetation is wilting and soil moisture availability is low.
3) MUREX
This case studied the land–atmosphere interaction over a C3 fallow field on an agricultural site near Toulouse, France, on 4 September 1997 (Calvet et al. 1999). Two grass species, Brachypodium sp. and Potentilla reptans, dominated the domain with a vegetation cover of 0.95 and an LAI of 2.5. The soil texture was silty-loam with surface albedo and emissivity values of 0.21 and 0.97. The model was integrated from 0000 LT 4 September 1997 for a period of 24 h. Figures 9a–c show the observed and the model-predicted radiative fluxes for the MUREX case. For the Rn, observations showed a positive value between 0600 and 1700 LT with a noontime peak around 600 W m−2. The simulated Rn values followed the observations closely, though the midday values were underestimated by ∼50 W m−2. This underestimation could have been due to a number of possible interactions, and our analysis suggests that the uncertainty in the albedo observations (by ±0.04) and heterogeneity in the fallow field were the main reasons for the differences in the model results (cf. Calvet et al. 1999). Barring this underestimation, the model-simulated values were within 10% of the observations.
Also, as shown in Fig. 9b, the simulated LHF is in agreement with the observations. This agreement held true especially from the afternoon through to the end of the simulation. This general agreement could be largely attributed to the availability of site-specific vegetation and surface characteristics (Dang et al. 1998). For the earlier part of the day, the SHF values were underpredicted, and the observations showed an unusual peak in SHF at 0900 LT. Corresponding to this underprediction in the mornings, the modeled SHF outcome was higher than the observations by about 40 W m−2 at midday. After 1500 LT, the simulated values followed the observations. The errors in the net radiation estimation could also have offset the SHF values.
c. Simulation over C3 forest
1) Forest-wet
The FOREST-WET and the FOREST-DRY cases validated the GEM-based SVAT over a C3 woody canopy in a forested area. The measurements were made over the Sandflats, a mixed forest of the Adirondack Mountain region of New York (Finkelstein et al. 2000). The site was instrumented with a 36-m tower, and the energy fluxes were measured at a height of approximately 10 m above the average tree cover using a sonic anemometer and an infrared gas analyzer. The vegetation canopy showed complex, inhomogeneous variations. Details regarding the instrumentation, the data quality control, and the seasonal changes in the site characteristics can be found in Finkelstein et al. (2000); the surface features are shown in Table 1. The model was initialized at 0000 LT 24 July 1998, using an Eta Model–based sounding centered over the study site. A complete diurnal cycle was simulated.
Figures 10a–c show the observed and model-simulated radiative flux. The large variations in the observed net radiation values are possibly due to the presence of scattered (subgrid scale) clouds and an undetermined forest energy storage term (McCaughey 1985). In the GEM, we adopted a big-leaf approach and did not consider multiple canopy layers (Meyers et al. 1998) or biomass energy storage (Gu et al. 2007), thus the variations due to within-canopy processes cannot be reproduced. The simulated net radiation values showed much of the features seen in the observed Rn, such as the peak value of 700 W m−2, although the observed variability in the midday values was absent.
Regarding the LHF and SHF, the observations showed variability that is not apparent in the Rn data and may be indicative of large within-canopy or surface storage flux. The fluxes varied by about 150 W m−2 between subsequent hourly readings. The LHF values were more variable than the relatively smooth Rn and SHF curves. The SHF values showed a smooth rise and peak around 150 W m−2 and then decreased to negative values after 1700 LT. The observations showed variability until noon and then peaked at about 200 W m−2 for a short period after fluctuating close to 0 and 100. Thus for this case, the SHF values were generally overpredicted. The model performance for this simulation was less satisfactory than it was for the grass and crops, though the general features were well simulated in terms of the evapotranspiration-based exchanges.
2) Forest-dry
The FOREST-DRY simulations were performed over the same forest canopy except the surface conditions were relatively drier. The model was initialized at 0000 LT 10 October 1998 and integrated for a period of 24 h. Figures 11a–c show the observed and modeled net radiation, latent heat fluxes, and sensible heat fluxes, respectively. The simulated values followed the net radiation observations until sunrise, after which the observed and simulated Rn values show noticeable differences. The peak Rn prediction was approximately 120 W m−2, and was consistent with observations, but the timings varied. There was also an overprediction in the simulated LHF. Even though the short-term variability was not captured, the model did perform well in simulating the case-to-case variability with generally lower LHF values for the FOREST-DRY case as compared with the FOREST-WET case.
Considering the variability in the observations and the difficulty associated with simulating the energy balance in patchy forest canopies, the model performance was modest. The results also suggest that the simplified scaling approach used in the land surface model could be applied for most mesoscale applications; however, for tall canopies, this approach could lead to larger uncertainties in the evapotranspiration and energy flux outcome. The scaling approach and the surface energy storage term for forested landscapes (Gu et al. 2007) need to be evaluated in a future study.
Recently, Kumar et al. (2007, 2008) and A. Kumar et al. (2008, unpublished manuscript) tested the GEM scheme within the NCAR high-resolution land data assimilation system with the Noah land surface model to simulate a month-long surface energy balance and soil moisture/soil temperature time series over the eastern half of the United States. The model results were compared with AmeriFlux observations (Baldocchi et al. 2001). Many of the validation sites were over forested landscapes and a very good agreement was seen, particularly over Morgan Forest in southern Indiana, Walker Branch in Tennessee, Duke Forest site in North Carolina, and Park Falls site in Wisconsin. Therefore, even though we obtained relatively modest results in this study with the 1D coupled modeling study for the FOREST-WET and FOREST-DRY cases, the viability of the GEM-based Rs and energy balance or soil fields can be considered good even over forest site.
d. GEM-predicted stomatal resistances
Based on the various field evaluations, we conclude that the default values used in the GEM-based coupled SVAT model were able to simulate the surface energy balance and the boundary layer feedback over different natural surfaces reasonably well (at least within the range of uncertainty associated with a coupled column model). An important variable that GEM simulates for the SVAT and that forms the basis for replacing the Jarvis-type transpiration model is stomatal resistance (Rs) [and carbon assimilation rate (An)]. Field measurements of Rs and An were available for the MUREX case. The MUREX observations and the corresponding model results are shown in Figs. 12a,b. Although there are apparent differences in the outcome, the variations are realistic considering the heterogeneity in the natural landscape as well as the intraspecies variations in the two dominant fallow grasses (Brachypodium sp. and Potentilla reptans) found on the observational site (Calvet et al. 1999). As discussed in Dang et al. (1998) and Colello et al. (1998), the performance of the gs–An models cannot be compared with every plant species that a particular vegetation type represents in a land surface model. Thus, from this perspective, similiarity of the observed and modeled gs–An terms is satisfactory.
To put the results in perspective, we evaluated the model response when the traditional minimum stomatal resistance (Rs,min) based Jarvis scheme was used instead of the GEM. For this testing we used the same surface conditions as used for GEM (shown in Table 1), except with the Jarvis-type model (Alapaty et al. 1997a). Note that the critical variable in adopting the Jarvis scheme is the specification of Rs,min. The vegetation type over the MUREX study area (natural fallow landscape) can be classified under a number of land use categories in a land surface model, such as the one adopted in the Noah land surface model (Ek et al. 2003). Considering the U.S. Geological Survey (USGS) land use category, one could prescribe the fallow landscape as grasslands with shrubs (Rs,min = 150 s m−1) or even possibly as an agricultural area (Rs,min = 40 s m−1). Other possibilities include considering the landscape as grassland (Rs,min = 40 s m−1), mixed shrub or grasslands (Rs,min = 170 s m−1), or shrubland (Rs,min = 300 s m−1). As discussed before, the Rs,min assignment is an arbitrary parameter used only in the models, and thus compiled from the literature and modeling studies, and does not have a measurement-based reference. Further, considering the broad definitions of each of the land use categories, no guiding principle exists for choosing one land use category over the other. Hence, corresponding to these different realistically possible vegetation-type assignments for the fallow surface, we prescribed three corresponding Rs,min values—40, 150, and 300 s m−1—and compared the model outcome for stomatal resistance (Rs) temporal changes.
Figure 13 shows the observed and modeled stomatal resistances over the fallow site. As discussed earlier, the GEM-based Rs values were close to the observations for the two dominant species over the study domain. The Jarvis-type outcome was of the same order of magnitude but showed a large dependence on the Rs,min assignment. The model performance depends on, and to a large extent can be tuned to, the prescribed value of minimum stomatal resistance (Niyogi and Raman 1997). Consequently, results for Rs,min = 40 s m−1 do yield stomatal resistance values relatively closer to the observed (though underpredicted), as compared with the high Rs values yielded by the other two Rs,min assignments. Thus, as shown in Alfieri et al. (2008), one could in practice tune the Rs,min value (in this particular case, increase it slightly) to achieve a good stomatal resistance variation from the Jarvis scheme. [Regarding the MUREX study referred to here, Calvet et al. (1998) used Rs,min = 50 s m−1 in one of their simulations and obtained fairly good agreement between the modeled and observed values.]
Changes in the Rs,min values affected the modeled LHF. Higher stomatal resistance resulted in lower latent heat fluxes (Fig. 14). The modeled LHF values during the afternoon ranged from about 150 to 350 W m−2 for the fallow land, while the observations showed values around 300 W m−2. Consistent with the earlier discussion, the GEM results were in close agreement with the LHF, and also with the Rs,min = 40 case results.
In addition to these direct observations available during MUREX, a number of studies such as Pleim and Xiu (1995) and Alapaty et al. (1997a,b) have calibrated the Jarvis-type Noilhan–Planton scheme (principally the Rs,min values) for integrated field campaigns. As discussed for the MUREX case study earlier, the Jarvis-type Rs time series can be considered a surrogate for observed values if the Rs,min is calibrated with other surface observations. Hence, we used the simulated stomatal resistance time series using the Jarvis scheme for the two FIFE cases: FIFE1 and FIFE2 following Alapaty et al. (1997a), to compare the GEM-predicted Rs values for the two cases. Figures 15a,b show the stomatal resistance variations from the two schemes over the FIFE region. In general, the Rs values are similar, as expected because of the close agreement between the observed and GEM-predicted surface energy fluxes shown earlier (Figs. 1, 3). The GEM predictions showed a somewhat larger Rs than the calibrated values, but the differences were relatively small (as revealed from comparisons with observed profiles and surface energy fluxes). Thus the photosynthesis/gas exchange–based evapotranspiration approach to estimating stomatal resistances (net carbon assimilation) as part of coupled surface–atmosphere models was shown in this study to be robust and effective.
5. Conclusions
In the majority of mesoscale weather forecast models, the Jarvis-type scheme is employed for simulating vegetation feedback in a land surface model. We showed that the Jarvis scheme, despite its robustness, needs significant tuning to simulate surface energy balances over different natural surfaces (Alfieri et al. 2008). Recent advances in vegetation models have developed explicit relations for modeling photosynthesis and stomatal conductance, and subsequently the variations in surface evapotranspiration. These photosynthesis or carbon assimilation–based stomatal models have been successfully employed at the leaf scale and in climate studies. In this study, we showed that a photosynthesis-based stomatal resistance model despite its complexity can efficiently replace the Jarvis scheme for coupled mesoscale land–atmosphere models. To show this, we developed and coupled a gas exchange–based surface evapotranspiration model (GEM) as a land surface/SVAT scheme designed for mesoscale applications.
The GEM was dynamically coupled with a prognostic soil moisture, soil temperature model, and a PBL model. This coupled system was then validated with nine case studies over five different natural surfaces: a C4 prairie grass, a C4 corn field, a C3 soybean field, a C3 fallow site, and a C3 hardwood forest site. For each of the five surfaces except the fallow site, two case studies were performed under contrasting surface conditions (such as different soil moisture and LAI). The model simulations were compared with actual field measurements of surface sensible and latent heat fluxes. For some of the case studies, measurements of vertical boundary layer profiles, or direct measurements of stomatal resistance and photosynthesis rates were available and were compared with the GEM-based, coupled SVAT model output.
The model was able to simulate the different surface as well as boundary layer characteristics quite successfully. The surface energy fluxes in general, and the average latent heat fluxes in particular, were within 10%–20% of the observations without any tuning of the biophysical–vegetation characteristics. The model was also able to satisfactorily simulate the day-to-day variations in the heat fluxes. The model response to the changes in the surface characteristics was consistent with observations and theory.
A number of features can be further developed in the parameterization. For example, in this study the big-leaf approach was adopted, which could have caused significant errors for tall canopies, particularly when a large storage flux may be involved. A future study could be undertaken redesigning the coupling with a detailed sun–shade scaling and a multilayer modeling approach, particularly for tall canopies. Additionally, the photosynthesis-based model performance was relatively poor under low–soil moisture conditions. It should be mentioned that for grass and homogeneous crop canopies, the big-leaf approach was satisfactory, particularly under high–soil moisture conditions or when the vegetation was not aged or stressed.
The case studies provide two significant conclusions: (i) the inclusion of a new photosynthesis-based vegetation scheme did not deteriorate the overall performance of the coupled model, and (ii) inclusion of the new photosynthesis transpiration scheme and the generally good results obtained without any significant tuning of the physiological variables in the vegetation model validated the robust coupling between GEM and the ABL. Results thus suggest that the photosynthesis-based SVAT approaches are superior to Jarvis-based approaches and can be applied to regional environmental models. Also, the GEM has been successfully coupled with the Noah land surface model within the Coupled Ocean–Atmospheric Modeling System (Niyogi et al. 2006; Holt et al. 2006) and the Weather Research and Forecasting system (Kumar et al. 2008; Chang et al. 2009); the results for the coupled 3D simulations as well as the seasonal to longer-term offline simulations show the GEM capable of the reproduction of a realistic land surface response (Kumar et al. 2008; A. Kumar et al. 2008, unpublished manuscript).
Acknowledgments
This study benefited in part from the following grants: NSF-ATM 0233780 (Dr. S. Nelson), NASA-THP NNG04GI84G and NNG06GH17G (Dr. J. Entin), NASA-IDS NNG04GL61G (Drs. J. Entin and G. Gutman), NASA Land Use Landcover Change Program (Dr. G. Gutman), USDA NRICGP on Water Resources (subcontract though Tufts University Dr. S. Islam), and the NOAA Joint Center for Satellite Data Assimilation NA06NES4400013 (Dr. K. Mitchell). We also acknowledge the support from the NCAR Water Cycle Across Scale Program at The Institute for Integrative and Multidisciplinary Earth Studies. We also appreciate detailed comments and suggestions from three reviewers.
REFERENCES
Alapaty, K., J. Pleim, S. Raman, D. Niyogi, and D. Byun, 1997a: Simulation of atmospheric boundary layer processes using local- and nonlocal-closure schemes. J. Appl. Meteor., 36 , 214–233.
Alapaty, K., S. Raman, and D. Niyogi, 1997b: Uncertainty in specification of surface characteristics: A study of prediction errors in the boundary layer. Bound.-Layer Meteor., 82 , 473–500.
Alapaty, K., N. Seaman, D. Niyogi, and A. Hanna, 2001: Assimilating surface data to improve the accuracy of atmospheric boundary layer simulations. J. Appl. Meteor., 40 , 2068–2082.
Alfieri, J. G., D. Niyogi, P. D. Blanken, F. Chen, M. A. LeMone, K. Mitchell, M. B. Ek, and A. Kumar, 2008: Estimation of the minimum canopy resistance for croplands and grasslands using data from the 2002 International H2O Project. Mon. Wea. Rev., 136 , 4452–4469.
Anderson, N., J. Norman, T. Meyers, and G. Diak, 2000: An analytical model for estimating canopy transpiration and carbon assimilation fluxes based on canopy light use efficiency. Agric. For. Meteor., 101 , 265–289.
Baldocchi, D., 1992: A Lagrangian random-walk model for simulating water vapor, CO2, and sensible heat flux densities and scalar profiles over and within a soybean canopy. Bound.-Layer Meteor., 61 , 113–144.
Baldocchi, D., and Coauthors, 2001: FLUXNET: A new tool to study the temporal and spatial variability of ecosystem–scale carbon dioxide, water vapor, and energy flux densities. Bull. Amer. Meteor. Soc., 82 , 2415–2434.
Ball, J., I. Woodrow, and J. Berry, 1987: A model predicting stomatal conductance and its contribution to the control of photosynthesis under different environmental conditions. Progress in Photosynthesis Research, Vol. 4, J. Biggins, Ed., Martinus Nijhoff, 221–224.
Blackadar, A., 1976: Modeling the nocturnal boundary layer. Proc. Third Symp. on Atmospheric Turbulence, Diffusion, and Air Quality, Raleigh, NC, Amer. Meteor. Soc., 46–49.
Bosilovich, M., and W. Sun, 1995: Formulation and verification of a land surface parameterization for atmospheric models. Bound.-Layer Meteor., 73 , 321–341.
Calvet, J-C., J. Noilhan, J. Roujean, P. Bessemoulin, M. Cabelguenne, A. Olioso, and J. Wigneron, 1998: An interactive vegetation SVAT model tested against data from six contrasting sites. Agric. For. Meteor., 92 , 73–95.
Calvet, J., and Coauthors, 1999: MUREX: A land surface field experiment to study the annual cycle of the energy and water budgets. Ann. Geophys., 17 , 838–854.
Chang, H., A. Kumar, D. Niyogi, U. C. Mohanty, F. Chen, and J. Dudhia, 2009: The role of land surface processes on the mesoscale simulation of the July 26, 2005 heavy rain event over Mumbai, India. Global Planet. Change, in press.
Chen, F., and J. Dudhia, 2001: Coupling an advanced land surface/hydrology model with the Penn State/NCAR MM5 modeling system. Part I: Model implementation and sensitivity. Mon. Wea. Rev., 129 , 569–585.
Chen, F., and Coauthors, 1996: Modeling of land surface evaporation by four schemes and comparison with FIFE observations. J. Geophys. Res., 101 , 7251–7268.
Clapp, R., and G. Hornberger, 1978: Empirical equations for some soil hydraulic properties. Water Resour. Res., 14 , 601–604.
Colello, G., C. Grivet, P. Sellers, and J. Berry, 1998: Modeling of energy, water, and CO2 flux in a temperate grassland ecosystem with SiB2: May–October 1987. J. Atmos. Sci., 55 , 1141–1169.
Collatz, J., J. Ball, C. Grivet, and J. Berry, 1991: Physiological and environmental regulation of stomatal conductance, photosynthesis and transpiration: A model that includes a laminar boundary layer. Agric. For. Meteor., 54 , 107–136.
Collatz, J., M. Ribas-Carbo, and J. Berry, 1992: Coupled photosynthesis-stomatal conductance model for leaves of C4 plants. Aust. J. Plant Physiol., 19 , 519–538.
Cooter, E., and D. Schwede, 2000: Sensitivity of the National Oceanic and Atmospheric Administration multilayer model to instrument error and parameterization uncertainty. J. Geophys. Res., 105 , 6695–6704.
Cowan, I., 1982: Regulation of water use in relation to carbon gain in higher plants. Physiological Plant Ecology II, O. Lange et al., Eds., Vol. 12B, Encylopedia of Plant Physiology, New Series, Springer-Verlag, 589–615.
Cox, P., R. Betts, C. Bunton, R. Essery, P. Rowntree, and J. Smith, 1999: The impact of new land surface physics on the GCM simulation of climate and climate sensitivity. Climate Dyn., 15 , 183–203.
Dai, Y., and Coauthors, 2003: The Common Land Model. Bull. Amer. Meteor. Soc., 84 , 1013–1023.
Dang, Q., H. Margolis, and J. Collatz, 1998: Parameterization and testing of a coupled photosynthesis-stomatal conductance model for boreal trees. Tree Physiol., 18 , 141–153.
Deardorff, J., 1978: Efficient prediction of ground surface temperature and moisture with inclusion of a layer of vegetation. J. Geophys. Res., 83 , 1889–1903.
Ek, M. B., K. E. Mitchell, Y. Lin, E. Rogers, P. Grunmann, V. Koren, G. Gayno, and J. D. Tarpley, 2003: Implementation of Noah land surface model advances in the National Centers for Environmental Prediction operational mesoscale Eta model. J. Geophys. Res., 108 , 8851. doi:10.1029/2002JD003296.
Farquhar, G., S. von Caemmerer, and J. Berry, 1980: A biochemical model of photosynthetic CO2 assimilation in leaves of C3 species. Planta, 149 , 78–90.
Feddema, J. J., K. W. Oleson, G. B. Bonan, L. O. Mearns, L. E. Buja, G. A. Meehl, and W. M. Washington, 2005: The importance of land-cover change in simulating future climates. Science, 310 , 1674–1678.
Finkelstein, P., T. Ellestad, J. Clarke, T. Meyers, D. Schwede, E. Hebert, and J. Neal, 2000: Ozone and sulfur dioxide dry deposition to forests: Observations and model evaluation. J. Geophys. Res., 105 , 15365–15378.
Goudriaan, J., H. van Laar, H. van Keulen, and W. Lovwerse, 1985: Photosynthesis, CO2 and plant production. Wheat Growth and Modeling, W. Day and R. K. Atkin, Eds., NATO ASI Series A, Vol. 86, Plenum Press, 107–122.
Gu, L., and Coauthors, 2007: Influences of biomass heat and biochemical energy storages on the land surface fluxes and radiative temperature. J. Geophys. Res., 112 , D02107. doi:10.1029/2006JD007425.
Holt, T., D. Niyogi, F. Chen, M. A. LeMone, K. Manning, and A. L. Qureshi, 2006: Effect of land–atmosphere interactions on the IHOP 24–25 May 2002 convection case. Mon. Wea. Rev., 134 , 113–133.
Idso, S. B., R. D. Jackson, R. J. Reginato, B. A. Kimball, and F. S. Nakayama, 1975: The dependence of bare soil albedo on soil water content. J. Appl. Meteor., 14 , 109–113.
Jacobs, C., 1994: Direct impact of atmospheric CO2 enrichment on regional transpiration. Ph.D. thesis, Wageningen Agricultural University, 179 pp.
Jarvis, P., 1976: The interpretation of leaf water potential and stomatal conductance found in canopies in the field. Philos. Trans. Roy. Soc. London, 273B , 593–610.
Jones, H. G., 1998: Stomatal control of photosynthesis and transpiration. J. Exp. Bot., 49 , 387–398.
Kim, J., and S. B. Verma, 1991: Modeling canopy photosynthesis: Scaling up from a leaf to canopy in a temperate grassland ecosystem. Agric. For. Meteor., 57 , 187–208.
Kumar, A., F. Chen, D. Niyogi, K. Manning, M. Ek, and K. Mitchell, 2007: Impact of canopy resistance formulation on the long-term evolution of soil moisture and evaporation. Preprints, 21st Conf. on Hydrology, San Antonio, TX, Amer. Meteor. Soc., 4.5. [Available online at http://ams.confex.com/ams/87ANNUAL/techprogram/paper_116815.htm.].
Kumar, A., F. Chen, D. Niyogi, K. Manning, M. Ek, K. Mitchell, and J. G. Alfieri, 2008: Integrating a photosynthesis-based canopy resistance model and new MODIS-based data to improve the presentation of vegetation transpiration in the Noah land surface model. Preprints, 22nd Conf. on Hydrology, New Orleans, LA, Amer. Meteor. Soc., J3.1. [Available online at http://ams.confex.com/ams/pdfpapers/131425.pdfm.].
LeMone, M. A., and Coauthors, 2007: NCAR/CU surface, soil, and vegetation observations during the International H20 Project 2002 Field Campaign. Bull. Amer. Meteor. Soc., 88 , 65–81.
LeMone, M. A., M. Tewari, F. Chen, J. G. Alfieri, and D. Niyogi, 2008: Evaluation of the Noah land surface model using data from a fair-weather IHOP_2002 day with heterogeneous surface fluxes. Mon. Wea. Rev., 136 , 4915–4941.
Leuning, R., 1990: Modeling stomatal behavior and photosynthesis of Eucalyptus grand. Aust. J. Plant Physiol., 17 , 150–175.
Maier-Maercker, U., 1998: Dynamics of change in stomatal response and water status of Picea abies during a persistent drought period: A contribution to the traditional view of plant water relations. Tree Physiol., 18 , 211–222.
Makela, A., F. Berninger, and P. Hari, 1996: Optimal control of gas exchange during drought: Theoretical analysis. Ann. Bot., 77 , 461–467.
Masson, V., J-L. Champeaux, F. Chauvin, C. Meriguet, and R. Lacaze, 2003: A global database of land surface parameters at 1-km resolution in meteorological and climate models. J. Climate, 16 , 1261–1282.
McCaughey, J. H., 1985: Energy balance storage terms in a mature mixed forest at Petawawa, Ontario—A case study. Bound.-Layer Meteor., 31 , 89–101. doi:10.1007/BF00120036.
Meyers, T., P. Finkelstein, J. Clarke, T. Ellestad, and P. Sims, 1998: A multilayer model for inferring dry deposition using standard meteorological measurements. J. Geophys. Res., 103 , 22645–22661.
Nikolov, N., W. Massman, and A. Schoettle, 1995: Coupling biochemical and biophysical processes at the leaf level: An equilibrium photosynthesis model for leaves of C3 plants. Ecol. Model., 80 , 205–235.
Niyogi, D., and S. Raman, 1997: Comparison of four different stomatal resistance schemes using FIFE data. J. Appl. Meteor., 36 , 903–917.
Niyogi, D., and S. Raman, 2001: Numerical modeling of gas deposition and bi-directional surface-atmosphere exchanges in mesoscale air pollution systems. Mesoscale Atmospheric Dispersion, E. Z. Boybeyi, Ed., Advances in Air Pollution, Vol. 9, WIT, 424–484.
Niyogi, D., S. Raman, A. Prabhu, U. Kumar, and S. S. Joshi, 1997: Direct estimation of stomatal resistance for meteorological applications. Geophys. Res. Lett., 24 , 1771–1774.
Niyogi, D., S. Raman, and K. Alapaty, 1998: Comparison of four different stomatal resistance schemes using FIFE observations. Part II: Analysis of terrestrial biospheric–atmospheric interactions. J. Appl. Meteor., 37 , 1301–1320.
Niyogi, D., S. Raman, and K. Alapaty, 1999: Uncertainty in the specification of surface characteristics, Part 2: Hierarchy of interaction explicit statistical analysis. Bound.-Layer Meteor., 91 , 341–366.
Niyogi, D., T. Holt, S. Zhong, P. C. Pyle, and J. Basara, 2006: Urban and land surface effects on the 30 July 2003 mesoscale convective system event observed in the Southern Great Plains. J. Geophys. Res., 111 , D19107. doi:10.1029/2005JD006746.
Noilhan, J., and S. Planton, 1989: A simple parameterization of land surface processes for meteorological models. Mon. Wea. Rev., 117 , 536–549.
Paw U, K., 1987: Mathematical analysis of the operative temperature and energy budget. J. Therm. Biol., 3 , 227–233.
Pielke Sr., R. A., and D. Niyogi, 2009: The role of landscape processes within the climate system. Landform—Structure, Evolution, Process Control: Proceedings of the International Symposium on Landforms Organised by the Research Training Group 437, Vol. 115, J. C. Otto and R. Dikaum, Eds., Springer, in press.
Pielke Sr., R. A., G. Marland, R. A. Betts, T. N. Chase, J. L. Eastman, J. O. Niles, D. Niyogi, and S. Running, 2002: The influence of land-use change and landscape dynamics on the climate system: Relevance to climate change policy beyond the radiative effect of greenhouse gases. Philos. Trans. Roy. Soc. London, 360A , 1705–1719.
Pitman, A. J., 2003: The evolution of, and revolution in, land surface schemes designed for climate models. Int. J. Climatol., 23 , 479–510. doi:10.1002/joc.893.
Pleim, J., and A. Xiu, 1995: Development and testing of a surface flux planetary boundary layer model with explicit soil moisture parameterization for applications in mesoscale models. J. Appl. Meteor., 34 , 16–32.
Polley, H., J. Norman, T. Arkebauer, E. Walter-Shea, and D. Greegor, 1992: Leaf-gas exchange of Andropogon Gerardii Vitman, Panicum Virgatum L., and Sorgastrum Nutans (L.) Nash in a Tallgrass prairie. J. Geophys. Res., 97 , 18837–18844.
Schulze, E., F. Kelliher, C. Korner, J. Lloyd, and R. Leuning, 1994: Relationships among maximum stomatal conductance, ecosystem surface conductance, carbon assimilation rate, and plant nitrogen nutrition: A global ecology scaling exercise. Annu. Rev. Ecol. Syst., 25 , 629–660.
Sellers, P., F. Hall, G. Asrar, D. Strebel, and R. Murphy, 1992: An overview of the First International Satellite Land Surface Climatology Project (ISLSCP) Field Experiment (FIFE). J. Geophys. Res., 97 , 18345–18371.
Sellers, P. J., and Coauthors, 1996: A revised land surface parameterization (SiB2) for atmospheric GCMs. Part I: Model formulation. J. Climate, 9 , 676–705.
Su, H., K. Paw U, and R. Shaw, 1996: Development of a coupled leaf and canopy model for the simulation of plant–atmosphere interaction. J. Appl. Meteor., 35 , 734–748.
van Heemst, H. D. J., 1988: Plant data values required for simple crop growth simulation models: Review and bibliography. Simulation Reps. CABO-TT 17, Centre for Agrobiological Research, Department of Theoretical Production Ecology, Agricultural University, 100 pp.
Wilson, K., D. Baldocchi, and P. Hanson, 2000a: Spatial and seasonal variability of photosynthetic parameters and their relationship to leaf nitrogen in a deciduous forest. Tree Physiol., 20 , 565–578.
Wilson, K., D. Baldocchi, and P. Hanson, 2000b: Quantifying stomatal and non stomatal limitations to carbon assimilation resulting from leaf aging and drought in mature deciduous tree species. Tree Physiol., 20 , 787–797.
Wilson, K., and Coauthors, 2002: Energy balance closure at FLUXNET sites. Agric. For. Meteor., 113 , 223–243.
APPENDIX A
The Photosynthesis/Carbon Assimilation Model
In the GEM, we couple the photosynthesis/carbon assimilation model developed by Collatz et al. (1991, 1992) with the Ball–Berry model [Eq. (1) in the main text]. Photosynthesis is calculated for three potentially limiting factors: (i) efficiency of the photosynthetic enzyme system or Rubisco limitation (wc); (ii) amount of photosynthetically active radiation (PAR) absorbed by the leaf chlorophyll or light limitation (we); and (iii) capacity of the C3 vegetation to utilize the photosynthesis products, or the phosphoenolpyruvate (PEP)–carboxylase limitation in the C4 vegetation (ws).
APPENDIX B
Leaf Boundary Layer Scheme for the Stomatal Conductance/Photosynthesis Model
Observed (open circle, dashed line) and GEM-predicted (solid line) (a) net radiation (W m−2), (b) LHF (W m−2), and (c) SHF (W m−2) for 6 Jun 1987 FIFE1 case. The vegetation type is a C4-type grassland.
Citation: Journal of Applied Meteorology and Climatology 48, 2; 10.1175/2008JAMC1662.1
Observed (open circle, dashed line) and GEM-predicted (solid line) (a) net radiation (W m−2), (b) LHF (W m−2), and (c) SHF (W m−2) for 6 Jun 1987 FIFE1 case. The vegetation type is a C4-type grassland.
Citation: Journal of Applied Meteorology and Climatology 48, 2; 10.1175/2008JAMC1662.1
Observed (open circle, dashed line) and GEM-predicted (solid line) (a) net radiation (W m−2), (b) LHF (W m−2), and (c) SHF (W m−2) for 6 Jun 1987 FIFE1 case. The vegetation type is a C4-type grassland.
Citation: Journal of Applied Meteorology and Climatology 48, 2; 10.1175/2008JAMC1662.1
Observed (open circle; PBL-Obs) and GEM-estimated (closed circles; PBLHT) PBL height for 6 Jun 1987 FIFE1 case. The vegetation type is a C4-type grassland.
Citation: Journal of Applied Meteorology and Climatology 48, 2; 10.1175/2008JAMC1662.1
Observed (open circle; PBL-Obs) and GEM-estimated (closed circles; PBLHT) PBL height for 6 Jun 1987 FIFE1 case. The vegetation type is a C4-type grassland.
Citation: Journal of Applied Meteorology and Climatology 48, 2; 10.1175/2008JAMC1662.1
Observed (open circle; PBL-Obs) and GEM-estimated (closed circles; PBLHT) PBL height for 6 Jun 1987 FIFE1 case. The vegetation type is a C4-type grassland.
Citation: Journal of Applied Meteorology and Climatology 48, 2; 10.1175/2008JAMC1662.1
Observed (open circle, dashed line) and GEM-predicted (solid line) (a) net radiation (W m−2), (b) LHF (W m−2), and (c) SHF (W m−2) for 11 Jul 1987 FIFE2 case. The vegetation type is a C4-type grassland. The vegetation is at peak green as compared with FIFE1 (Fig. 2).
Citation: Journal of Applied Meteorology and Climatology 48, 2; 10.1175/2008JAMC1662.1
Observed (open circle, dashed line) and GEM-predicted (solid line) (a) net radiation (W m−2), (b) LHF (W m−2), and (c) SHF (W m−2) for 11 Jul 1987 FIFE2 case. The vegetation type is a C4-type grassland. The vegetation is at peak green as compared with FIFE1 (Fig. 2).
Citation: Journal of Applied Meteorology and Climatology 48, 2; 10.1175/2008JAMC1662.1
Observed (open circle, dashed line) and GEM-predicted (solid line) (a) net radiation (W m−2), (b) LHF (W m−2), and (c) SHF (W m−2) for 11 Jul 1987 FIFE2 case. The vegetation type is a C4-type grassland. The vegetation is at peak green as compared with FIFE1 (Fig. 2).
Citation: Journal of Applied Meteorology and Climatology 48, 2; 10.1175/2008JAMC1662.1
Observed (open circle; PBL-Obs) and GEM-estimated (closed circles; PBLHT) PBL height for 11 Jul 1987 FIFE1 case. The vegetation type is a C4-type grassland.
Citation: Journal of Applied Meteorology and Climatology 48, 2; 10.1175/2008JAMC1662.1
Observed (open circle; PBL-Obs) and GEM-estimated (closed circles; PBLHT) PBL height for 11 Jul 1987 FIFE1 case. The vegetation type is a C4-type grassland.
Citation: Journal of Applied Meteorology and Climatology 48, 2; 10.1175/2008JAMC1662.1
Observed (open circle; PBL-Obs) and GEM-estimated (closed circles; PBLHT) PBL height for 11 Jul 1987 FIFE1 case. The vegetation type is a C4-type grassland.
Citation: Journal of Applied Meteorology and Climatology 48, 2; 10.1175/2008JAMC1662.1
Observed (open circle, dashed line) and GEM-predicted (solid line) (a) net radiation (W m−2), (b) LHF (W m−2), and (c) SHF (W m−2) for 24–25 Aug 1994 CORN-HIGH case. The vegetation type is a fully grown corn crop.
Citation: Journal of Applied Meteorology and Climatology 48, 2; 10.1175/2008JAMC1662.1
Observed (open circle, dashed line) and GEM-predicted (solid line) (a) net radiation (W m−2), (b) LHF (W m−2), and (c) SHF (W m−2) for 24–25 Aug 1994 CORN-HIGH case. The vegetation type is a fully grown corn crop.
Citation: Journal of Applied Meteorology and Climatology 48, 2; 10.1175/2008JAMC1662.1
Observed (open circle, dashed line) and GEM-predicted (solid line) (a) net radiation (W m−2), (b) LHF (W m−2), and (c) SHF (W m−2) for 24–25 Aug 1994 CORN-HIGH case. The vegetation type is a fully grown corn crop.
Citation: Journal of Applied Meteorology and Climatology 48, 2; 10.1175/2008JAMC1662.1
Observed (open circle, dashed line) and GEM-predicted (solid line) (a) net radiation (W m−2), (b) LHF (W m−2), and (c) SHF (W m−2) for 15–16 Sep 1994 CORN-OLD case. The vegetation is an old corn crop.
Citation: Journal of Applied Meteorology and Climatology 48, 2; 10.1175/2008JAMC1662.1
Observed (open circle, dashed line) and GEM-predicted (solid line) (a) net radiation (W m−2), (b) LHF (W m−2), and (c) SHF (W m−2) for 15–16 Sep 1994 CORN-OLD case. The vegetation is an old corn crop.
Citation: Journal of Applied Meteorology and Climatology 48, 2; 10.1175/2008JAMC1662.1
Observed (open circle, dashed line) and GEM-predicted (solid line) (a) net radiation (W m−2), (b) LHF (W m−2), and (c) SHF (W m−2) for 15–16 Sep 1994 CORN-OLD case. The vegetation is an old corn crop.
Citation: Journal of Applied Meteorology and Climatology 48, 2; 10.1175/2008JAMC1662.1
Observed (open circle, dashed line) and GEM-predicted (solid line) (a) net radiation (W m−2), (b) LHF (W m−2), and (c) SHF (W m−2) for 6–7 Aug 1995 SOY-FULL case. The vegetation type is a fully grown soybean crop.
Citation: Journal of Applied Meteorology and Climatology 48, 2; 10.1175/2008JAMC1662.1
Observed (open circle, dashed line) and GEM-predicted (solid line) (a) net radiation (W m−2), (b) LHF (W m−2), and (c) SHF (W m−2) for 6–7 Aug 1995 SOY-FULL case. The vegetation type is a fully grown soybean crop.
Citation: Journal of Applied Meteorology and Climatology 48, 2; 10.1175/2008JAMC1662.1
Observed (open circle, dashed line) and GEM-predicted (solid line) (a) net radiation (W m−2), (b) LHF (W m−2), and (c) SHF (W m−2) for 6–7 Aug 1995 SOY-FULL case. The vegetation type is a fully grown soybean crop.
Citation: Journal of Applied Meteorology and Climatology 48, 2; 10.1175/2008JAMC1662.1
Observed (open circle, dashed line) and GEM-predicted (solid line) (a) net radiation (W m−2), (b) LHF (W m−2), and (c) SHF (W m−2) for 12–13 Sep 1995 SOY-FULL case. The vegetation type is an aged soybean crop.
Citation: Journal of Applied Meteorology and Climatology 48, 2; 10.1175/2008JAMC1662.1
Observed (open circle, dashed line) and GEM-predicted (solid line) (a) net radiation (W m−2), (b) LHF (W m−2), and (c) SHF (W m−2) for 12–13 Sep 1995 SOY-FULL case. The vegetation type is an aged soybean crop.
Citation: Journal of Applied Meteorology and Climatology 48, 2; 10.1175/2008JAMC1662.1
Observed (open circle, dashed line) and GEM-predicted (solid line) (a) net radiation (W m−2), (b) LHF (W m−2), and (c) SHF (W m−2) for 12–13 Sep 1995 SOY-FULL case. The vegetation type is an aged soybean crop.
Citation: Journal of Applied Meteorology and Climatology 48, 2; 10.1175/2008JAMC1662.1
Observed (open circle, dashed line) and GEM-predicted (solid line) (a) net radiation (W m−2), (b) LHF (W m−2), and (c) SHF (W m−2) for 4 Sep 1997 MUREX case. The vegetation type is a C3 fallow grassland.
Citation: Journal of Applied Meteorology and Climatology 48, 2; 10.1175/2008JAMC1662.1
Observed (open circle, dashed line) and GEM-predicted (solid line) (a) net radiation (W m−2), (b) LHF (W m−2), and (c) SHF (W m−2) for 4 Sep 1997 MUREX case. The vegetation type is a C3 fallow grassland.
Citation: Journal of Applied Meteorology and Climatology 48, 2; 10.1175/2008JAMC1662.1
Observed (open circle, dashed line) and GEM-predicted (solid line) (a) net radiation (W m−2), (b) LHF (W m−2), and (c) SHF (W m−2) for 4 Sep 1997 MUREX case. The vegetation type is a C3 fallow grassland.
Citation: Journal of Applied Meteorology and Climatology 48, 2; 10.1175/2008JAMC1662.1
Observed (open circle) and GEM-predicted (solid line) (a) net radiation (W m−2), (b) LHF (W m−2), and (c) SHF (W m−2) for 24 Jul 1998 FOREST-WET case.
Citation: Journal of Applied Meteorology and Climatology 48, 2; 10.1175/2008JAMC1662.1
Observed (open circle) and GEM-predicted (solid line) (a) net radiation (W m−2), (b) LHF (W m−2), and (c) SHF (W m−2) for 24 Jul 1998 FOREST-WET case.
Citation: Journal of Applied Meteorology and Climatology 48, 2; 10.1175/2008JAMC1662.1
Observed (open circle) and GEM-predicted (solid line) (a) net radiation (W m−2), (b) LHF (W m−2), and (c) SHF (W m−2) for 24 Jul 1998 FOREST-WET case.
Citation: Journal of Applied Meteorology and Climatology 48, 2; 10.1175/2008JAMC1662.1
Observed (open circle) and GEM-predicted (solid line) (a) net radiation (W m−2), (b) LHF (W m−2), and (c) SHF (W m−2) for 10 Oct 1998 FOREST-DRY case.
Citation: Journal of Applied Meteorology and Climatology 48, 2; 10.1175/2008JAMC1662.1
Observed (open circle) and GEM-predicted (solid line) (a) net radiation (W m−2), (b) LHF (W m−2), and (c) SHF (W m−2) for 10 Oct 1998 FOREST-DRY case.
Citation: Journal of Applied Meteorology and Climatology 48, 2; 10.1175/2008JAMC1662.1
Observed (open circle) and GEM-predicted (solid line) (a) net radiation (W m−2), (b) LHF (W m−2), and (c) SHF (W m−2) for 10 Oct 1998 FOREST-DRY case.
Citation: Journal of Applied Meteorology and Climatology 48, 2; 10.1175/2008JAMC1662.1
(a) Observed (dashed line) and GEM-estimated (solid line) net carbon assimilation (μmol m−2 s−1) for the MUREX case. (b) Modeled (solid line) and observed (dashed lines) stomatal resistances for the same MUREX case. Rs-brach and Rs-pot refer to the stomatal resistance observations for two dominant vegetation (same biome type) species: Brachypodium sp. and Potentilla reptans over the study site.
Citation: Journal of Applied Meteorology and Climatology 48, 2; 10.1175/2008JAMC1662.1
(a) Observed (dashed line) and GEM-estimated (solid line) net carbon assimilation (μmol m−2 s−1) for the MUREX case. (b) Modeled (solid line) and observed (dashed lines) stomatal resistances for the same MUREX case. Rs-brach and Rs-pot refer to the stomatal resistance observations for two dominant vegetation (same biome type) species: Brachypodium sp. and Potentilla reptans over the study site.
Citation: Journal of Applied Meteorology and Climatology 48, 2; 10.1175/2008JAMC1662.1
(a) Observed (dashed line) and GEM-estimated (solid line) net carbon assimilation (μmol m−2 s−1) for the MUREX case. (b) Modeled (solid line) and observed (dashed lines) stomatal resistances for the same MUREX case. Rs-brach and Rs-pot refer to the stomatal resistance observations for two dominant vegetation (same biome type) species: Brachypodium sp. and Potentilla reptans over the study site.
Citation: Journal of Applied Meteorology and Climatology 48, 2; 10.1175/2008JAMC1662.1
Stomatal resistance variations for the MUREX case, comparing the observations (Rs-brach and Rs-pot) with GEM predictions (Rc) and Jarvis-type predictions for 40 (Rs-40), 150 (Rs-150), and 300 s m−1 (Rs-300) as possible minimum stomatal resistance assignments for the fallow landscape.
Citation: Journal of Applied Meteorology and Climatology 48, 2; 10.1175/2008JAMC1662.1
Stomatal resistance variations for the MUREX case, comparing the observations (Rs-brach and Rs-pot) with GEM predictions (Rc) and Jarvis-type predictions for 40 (Rs-40), 150 (Rs-150), and 300 s m−1 (Rs-300) as possible minimum stomatal resistance assignments for the fallow landscape.
Citation: Journal of Applied Meteorology and Climatology 48, 2; 10.1175/2008JAMC1662.1
Stomatal resistance variations for the MUREX case, comparing the observations (Rs-brach and Rs-pot) with GEM predictions (Rc) and Jarvis-type predictions for 40 (Rs-40), 150 (Rs-150), and 300 s m−1 (Rs-300) as possible minimum stomatal resistance assignments for the fallow landscape.
Citation: Journal of Applied Meteorology and Climatology 48, 2; 10.1175/2008JAMC1662.1
As in Fig. 13 but for LHF variations.
Citation: Journal of Applied Meteorology and Climatology 48, 2; 10.1175/2008JAMC1662.1
As in Fig. 13 but for LHF variations.
Citation: Journal of Applied Meteorology and Climatology 48, 2; 10.1175/2008JAMC1662.1
As in Fig. 13 but for LHF variations.
Citation: Journal of Applied Meteorology and Climatology 48, 2; 10.1175/2008JAMC1662.1
GEM-estimated (soil line) and Jarvis-type calibrated (dashed line) stomatal resistances (s m−1) for (a) 6 Jun 1987 FIFE1 case and (b) 11 Jul 1987 FIFE2 case. Considering the uncertainty of the stomatal resistance estimation, the values are similar. The GEM values are based on default vegetation-type coefficients. The Jarvis scheme uses a calibrated minimum stomatal resistance from previously published studies.
Citation: Journal of Applied Meteorology and Climatology 48, 2; 10.1175/2008JAMC1662.1
GEM-estimated (soil line) and Jarvis-type calibrated (dashed line) stomatal resistances (s m−1) for (a) 6 Jun 1987 FIFE1 case and (b) 11 Jul 1987 FIFE2 case. Considering the uncertainty of the stomatal resistance estimation, the values are similar. The GEM values are based on default vegetation-type coefficients. The Jarvis scheme uses a calibrated minimum stomatal resistance from previously published studies.
Citation: Journal of Applied Meteorology and Climatology 48, 2; 10.1175/2008JAMC1662.1
GEM-estimated (soil line) and Jarvis-type calibrated (dashed line) stomatal resistances (s m−1) for (a) 6 Jun 1987 FIFE1 case and (b) 11 Jul 1987 FIFE2 case. Considering the uncertainty of the stomatal resistance estimation, the values are similar. The GEM values are based on default vegetation-type coefficients. The Jarvis scheme uses a calibrated minimum stomatal resistance from previously published studies.
Citation: Journal of Applied Meteorology and Climatology 48, 2; 10.1175/2008JAMC1662.1
Fig. B1. Flowchart showing GEM formulations.
Citation: Journal of Applied Meteorology and Climatology 48, 2; 10.1175/2008JAMC1662.1
Fig. B1. Flowchart showing GEM formulations.
Citation: Journal of Applied Meteorology and Climatology 48, 2; 10.1175/2008JAMC1662.1
Fig. B1. Flowchart showing GEM formulations.
Citation: Journal of Applied Meteorology and Climatology 48, 2; 10.1175/2008JAMC1662.1
Initial values of surface parameters used in the numerical simulations (Pleim and Xiu 1995; Alapaty et al. 1997a,b; Calvet et al. 1998; Meyers et al. 1998; Finkelstein et al. 2000).
