1. Introduction
Arid and semiarid regions are an important portion of the global land surface. Many studies have indicated that desertification is increasing as a result of climatic change and human activities (Puigdefabregas 1995; Warren 1996). The arid and semiarid region of northwestern China has experienced significant environmental changes within the last half century (Ma and Fu 2006). Thus, it is crucial to understand the land–atmosphere interactions and to predict the variations of the hydrometeorological regimes in these regions. From a lack of precipitation, the surface heat transport becomes the dominant land surface process (see section 3b) in arid regions. Therefore, parameterizing the surface heat transport process in land surface models (LSMs) is vital for accurately modeling the surface energy budget.
However, the ability of current LSMs to simulate land processes in arid and semiarid regions still needs improvement. Hogue et al. (2005) found that the Noah LSM tended to overestimate the sensible heat flux (H) and underestimate the surface temperature (Tsfc) during the dry season. Yang et al. (2007) evaluated seven general circulation models against coordinated enhanced observing period (CEOP) observations and found that all of the models significantly underestimated the daytime ground–air temperature differences, particularly severely in arid and semiarid regions. LeMone et al. (2008) showed that the default Noah model tended to overestimate H and underestimate Tsfc in relatively dry conditions. Yang et al. (2009) further evaluated three offline LSMs against observations in the area of the Tibetan Plateau (TP); their study indicated that all models significantly underestimated the daytime Tsfc. Chen and Zhang (2009) also found that the Noah model often overestimated the surface exchange coefficient for heat (Ch) over short vegetation.
The aforementioned modeling biases in Tsfc and H imply that the heat transfer resistance in the models is not appropriately parameterized. This resistance is related to both the aerodynamic roughness length (z0m) and the thermal roughness length (z0h). Much of the literature has focused on the parameterization of z0h (Sheppard 1958; Brutsaert 1982; Zilitinkevich 1995; Zeng and Dickinson 1998; Kanda et al. 2007; Smeets and van den Broeke 2008; Yang et al. 2008). Ma et al. (2002) and Yang et al. (2003) found that z0h exhibits an evident diurnal variation on the TP. Yang et al. (2008) evaluated several schemes and indicated that z0h depends on the flow state and exhibits diurnal variations. Some of these z0h schemes were developed and evaluated against observations in the micrometeorology community, while they need more critical evaluations and practical tests in LSMs to verify their effectiveness. That is the motivation for this study.
The objective of this work is twofold: 1) to assess the sensitivity of the land surface energy budget to different parameterizations of z0h in the Noah LSM and 2) to evaluate the performance of a revised Noah LSM against field observations in arid regions. The Noah LSM is selected because it is widely used and has been adopted for operations and research in National Centers for Environmental Prediction (NCEP) weather and climate predictions models and relevant data assimilation systems.
In this paper, section 2 briefly introduces the climatic characteristics of the sites and their relevant measurements. Section 3 describes the LSM used in this work and the settings of the model parameters. Section 4 tests the sensitivity to key LSM parameters in arid regions, and section 5 evaluates the performance of the revised LSM. Concluding remarks are given in section 6.
2. Sites and observation data
In this study, simulations are conducted at one semiarid site in Arizona and three arid sites in western China: two on the TP and one in the northwest. Some general information about these sites is given in Table 1.
The semiarid site is Audubon Research Ranch (referred to as Audubon hereafter), which is located in Arizona. The mean annual precipitation at Audubon is ∼300–400 mm. The land surface is characterized by sparse brown grass during the simulation period. The data were collected through the AmeriFlux network (information online at http://public.ornl.gov/ameriflux/). The required forcing data were measured by automatic weather stations. The observed ground truth data included the surface temperature, the soil temperature profile (2, 4, 8, 16, 32, 64, and 128 cm), the soil moisture profile (10, 20, 30, 40, 60, and 100 cm), and turbulent fluxes. The surface temperature was given by an infrared thermometer [Apogee Infrared Thermocouple Sensor (IRTS-P), Campbell Scientific]; the soil moisture sensor (PR1/6, Delta-T Devices) and soil temperature probe with thermistors (YSI Inc.) were set up to measure the soil moisture and soil temperature, respectively; and the turbulent fluxes was measured by an eddy-covariance system (LI-COR LI-7500 for carbon dioxide and water vapor concentrations, R.M. Young 81000V for wind speed and sonic temperature). The 30-min averages were recorded for all of the measurements. At this site, the simulation period is from 15 April to 1 June 2003, when no precipitation events occurred.
The two TP sites are Shiquanhe and Gaize, both located in the western TP with elevations ∼4000 m above sea level. Both sites, located in the midlatitude westerlies, belong to the alpine desert climate. The mean annual precipitation is around 200 mm and the land surface is almost bare soil. Owing to the high elevation and strong solar radiation, the surface heat fluxes and near-surface meteorological variables undergo especially evident diurnal variations. The measurements were collected through the Global Energy and Water Cycle Experiment (GEWEX) Asian Monsoon Experiment-Tibet (GAME-Tibet; Koike et al. 1999) during an intensive observing period (IOP, May∼September 1998). The forcing data were recorded by automatic weather stations. The observed ground truth data included surface temperature, soil temperature profile (5, 10, 20, 40, and 80 cm), and soil moisture at 0–15 cm; turbulent fluxes, however, were not available. The surface temperature was directly measured using a thermometer, with half of the sensor buried in the soil and half exposed to the air. This technique is used routinely by the Chinese Meteorological Administration to measure the bare soil surface temperature. The surface temperature measured using this technique agrees with that converted from the measured longwave radiation (given a surface emissivity of 0.9, which is derived by assuming the thermometer measurements near sunset are reliable) at Shiquanhe, with an uncertainty of 2–3 K. The soil moisture was measured with a time domain reflectometry (TDR) soil water content hygrometer. Data averaged over 30 or 60 min were recorded. At the Shiquanhe site, the simulation period is from 1 May to 14 September 1998, when the amount of rainfall is only 25 mm. At the Gerze site, the simulation period is shorter (from 1 May to 15 June 1998), though no precipitation events occurred.
The last arid desert site is Dunhuang. The observed data were collected through the Field Experiment on Interaction between Land and Atmosphere in Arid Region of North-west China (NWC-ALIEX; Zhang et al. 2005). Wind speed, air temperature, and humidity were measured on a tower at four heights: 1, 2, 8, and 18 m. Surface temperature was measured using the same technique as was used at Shiquanhe and Gaize. Soil temperatures were measured at six depths: 5, 10, 20, 40, 80, and 180 cm. Soil water contents were measured at depths of 5, 10, 20, and 80 cm with TDR. Downward and upward radiation flux components were measured at 1.5 m. During the IOP (10∼25 June 2004), turbulent fluxes were measured by the eddy-covariance technique with 30-min bulk averaging. The simulation targets the period 18 May–25 June, when no precipitation was detected.
3. Land surface model and parameter settings
a. Model description
The Noah LSM is developed based on the Oregon State University (OSU) LSM, which includes a two-layer soil model with thermal conduction equations for soil temperature and the diffusive form of Richardson’s equation for soil moisture (Mahrt and Pan 1984), as well as a Penman approach for the calculation of the latent heat flux (Mahrt and Ek 1984). After being widely adopted by NCEP, the Noah model has benefitted from a series of improvements. Among the major improvements are an increase from two to four soil layers, modifications to the canopy conductance formulation (Chen et al. 1996), bare soil evaporation and vegetation phenology (Betts et al. 1997), a new runoff formulation and infiltration (Schaake et al. 1996), thermal roughness length treatment in the surface layer exchange coefficients (Chen et al. 1997), and the inclusion of cold season processes (Koren et al. 1999). A more detailed overview of the physics lineage of the Noah LSM is presented in Ek et al. (2003).
We presently employ version 2.7 of the Noah LSM. In general, the model has four soil layers (with depths of 10, 30, 60, and 100 cm from top to bottom), a single canopy layer, and a single snow layer. The vegetation types are defined according to the categories assigned from U.S. Geological Survey (USGS) database. The soil types are defined by the Food and Agriculture Organization (FAO) database. Soil moisture for each soil layer is calculated from the diffusive form of Richard’s equation. Soil temperature is calculated from the heat diffusion equation. The surface temperature is determined following Mahrt and Ek (1984) to reflect a linearly combined ground–vegetation surface. A more detailed description of the model governing equations and the parameterizations can be found in Chen and Dudhia (2001).
b. Land surface processes in arid regions
The turbulent sensible heat and ground heat fluxes are the dominant components in Eq. (1b). In addition to the input forcing data, the parameters in the radiation and energy budgets (1a) and (1b) would determine the surface energy flux partitioning and the surface temperature. Therefore, they should be set accurately.
c. Model settings
The vegetation types at four sites are prescribed as the bare soil according to the aforementioned characteristics of the ground. The soil type is derived from the FAO data. The soil hydraulic parameters are much less important for the modeling of the surface temperature and energy budget at dry sites than are the energy-related soil parameters and surface parameters (α, ɛ, z0h, z0m, and kT). Among them, α, ɛ, and kT at all sites and z0m at Dunhuang can be derived from the observations, and their mean values are given in Table 2. The parameter kT was derived from the thermal diffusivity, which was estimated from the diurnal range of observed soil temperatures profile instead of a parameterization; α was directly obtained from observed downward and upward shortwave radiation fluxes; and ɛ was derived from surface temperature and longwave radiation fluxes.
In addition, z0m is physically related to the geometric roughness of surface elements and can be derived from the wind speed and temperature profiles. An optimal method suggested by Yang et al. (2008) was employed to estimate z0m at Dunhuang, where the profile data are available. The default value of z0m prescribed by vegetation type was used at the other three sites. In the Noah LSM, z0h is calculated by the Reynolds number–dependent scheme of Zilitinkevich (1995), as shown in Table 3. Yang et al. (2008) argued that this scheme overestimated z0h and thus underestimated the peak values of Tsfc. However, considering the significance of z0h and z0m, the sensitivity of thee surface energy budget to them will be investigated in section 4. In all simulations, the soil moisture and soil temperature are initialized with the observations.
4. Sensitivity test to the roughness lengths
Given the similarity among four arid sites, we regard Shiquanhe as being representative of the arid sites for a sensitivity analysis.
a. Thermal roughness length
To test the sensitivity of surface energy budget to different z0h values, six z0h schemes available in the literature were implemented into the Noah LSM. Listed in Table 3 are S58 (Sheppard 1958), B82 (Brutsaert 1982), Z95 (Zilitinkevich 1995), Z98 (Zeng and Dickinson 1998), K07 (Kanda et al. 2007), and Y08 (Yang et al. 2008). S58 and B82 were examined in detail in Verhoef et al. (1997). Z95 has been widely used in NCEP operational prediction systems since Chen et al. (1997). Z98 has been used in an LSM to unify undercanopy heat transfer processes between dense and sparse canopies. K07 was derived from urban canopy experiments. Y08 will be introduced in section 5a. In Yang et al. (2008), these schemes have been evaluated within the framework of Monin–Obukhov similarity theory by using observed Tsfc to parameterize H. In this study, their effectiveness is tested against independent datasets within the framework of land surface modeling.
Figure 1a compares the diurnal variation of Ch obtained using six schemes at Shiquanhe. During the simulation period, the highest Ch values of Z95 are about four times the lowest values of B82, the Ch values of Y08 are comparable to those of S58, and the results of K07 are very close to those of B82. Apparently, as in the previous studies, Ch is very sensitive to the parameterization schemes of z0h in arid regions.
In Figs. 1b and 1c, we compare dominant heat flux components simulated by different z0h schemes. As expected, differences in the simulated sensible heat flux arise from those in Ch values; high (low) Ch values correspond to high (low) H, and low (high) G0 are found accordingly. It is easy to interpret this phenomenon. As discussed in section 3, G0 is mainly determined by the modeled Tsfc. Therefore, if H is overestimated, Tsfc, which is calculated diagnostically from surface energy balance considerations, will be underestimated and thus G0 will be underestimated.
Figure 1d compares the diurnal variation of Tsfc between the observations and simulations with different z0h schemes. The simulations with the Y08 and S58 schemes produce good agreement with the observed daytime surface temperature, while the simulations with other schemes produce clear biases. Table 4 gives the comparative statistics between the observed Tsfc and the simulations using various schemes; we thus confirm that Y08 reproduces Tsfc more consistently with observed data than do the other schemes in this case.
b. Aerodynamic roughness length
The aerodynamic roughness length is ideally determined from the wind speed profile, although there have been some successes in relating this height to the arrangement, spacing, and physical height of individual roughness elements. The lack of profile data makes it hard to precisely estimate z0m. The common approach is to empirically prescribe a value of z0m for a given vegetation type.
The sensitivity of H and Tsfc to z0m is also investigated. In the simulations, we tried out two z0m values: a default value of z0m (0.011 m) and a lower value (0.001 m) at Shiquanhe, in combination with different z0h schemes in Table 3. Table 5 shows error metrics between the simulations produced by the six schemes. Clearly, different z0m values produce minor differences both in Tsfc and H; H and Tsfc, therefore, are not reasonably sensitive to z0m. Moreover, the Y08 scheme produces lower differences, indicating its particularly low sensitivity to the choice of z0m.
In summary, the surface flux and temperature simulations are highly sensitive to the z0h schemes and much less sensitive to the z0m value; the Y08 scheme seems to be a promising scheme since it can appropriately reproduce Tsfc. So, we applied the Y08 scheme to update the Noah LSM, and made further evaluations against the observations.
5. Updating Noah with the Y08 scheme and evaluations
a. Brief introduction to the Y08 scheme
The parameterization of the thermal surface roughness length is crucial for directly using Tsfc to calculate H. Many works (see Table 3) have related z0h directly to z0m through the parameter kB−1 [defined as ln(z0m/z0h)]. Momentum transport is generally more efficient than heat transport, due to the influence of pressure fluctuation, because individual roughness elements may enhance the momentum flux through form drag with little contribution to the area-averaged heat flux (Mahrt 1996). Therefore, z0h is typically less than z0m, especially over a surface with bluff roughness elements, and a higher z0m usually corresponds to a lower z0h.
b. Evaluations
Evaluations of both the revised Noah LSM and the original version were performed first at Audubon and Dunhuang, where surface temperature and sensible heat flux data are available. Then, they were evaluated at Shiquanhe and Gaize, where only surface temperature data are available.
1) Audubon and Dunhuang
Figures 2a–d compare the simulated diurnal variations of Tsfc, Rnet, H, and G0, respectively, between the revised Noah LSM and the original version at Audubon Research Ranch. Obviously, Tsfc and H, as well as Rnet, were properly simulated by the revised model, while the original model produced higher H and lower Tsfc, and, thus, higher Rnet. Table 6 gives error indices, which indicate that the revised model significantly reduced the simulation errors. The error metrics for Tsfc, Rnet, and H are calculated using data during the whole simulation period.
The major difference between the two simulations occurs in the daytime. Table 7 shows the error metrics using data during 0900–1600 local time (LT) when the original model yields higher H, higher Rnet, and lower Tsfc. The lower Tsfc would directly result in lower G0 (see Fig. 2d), corresponding to higher H.
Figures 3a–d compare the diurnal variations of the components in the energy balance equation at Dunhuang. As shown in Figs. 3a–c, the revised Noah LSM can better simulate Tsfc, Rnet, and H than could the original version. This is confirmed by the comparative statistics in Tables 6 and 7.
2) Shiquanhe and Gaize
Figures 4 and 5 give the comparisons at Shiquanhe and Gaize, respectively. Evidently, the original model significantly underestimated Tsfc and overestimated Rnet, while the revised model simulated well both Tsfc and Rnet. These results are confirmed by the error metrics in Table 6. The improvements by the revised model are especially evident in the daytime. In Table 7 we find that the original model underestimates Tsfc by more than 10 K during 0900–1600 LT, while the revised model performs much better. Moreover, we found that the original model performed worse at two TP sites than at Audubon and Dunhuang; this finding will be discussed in more depth in section 5c.
c. Discussion
To understand the different levels of performance between the revised Noah LSM and the original model, Fig. 6 shows the simulated mean diurnal variation of ln(z0h) throughout the simulation period at four sites. It is clear that the values of z0h simulated by the revised model exhibit evident diurnal variations. In fact, several studies have reported the diurnal variations of z0h over the bare-soil surface and grasslands (Verhoef et al. 1997; Sun 1999; Ma et al. 2002; Yang et al. 2003). But this diurnal variation was hardly simulated by the original Noah LSM. Figure 6 shows that the original model produces rather high z0h, and thus fairly high Ch, compared to those produced by the revised model in the daytime. Moreover, the originally simulated Tsfc is lower because the too high Ch carries too much heat away from the surface. Two direct effects arise from the underestimate of Tsfc: one is the overestimate of Rnet due to the reduced upward longwave radiation flux; the other is the underestimate of G0 due to an underestimated soil temperature gradient.
Figure 6 also shows that the values of z0h produced by the revised model have larger diurnal ranges at two TP sites than elsewhere. This is consistent with Yang et al. (2008), who found that the diurnal variations of z0h at TP sites are more evident than at other sites. This phenomenon may be attributed to strong diurnal changes of near-surface meteorological variables and higher land–atmosphere temperature differences caused by the high elevation and thus strong solar radiation. Because the differences between the two simulated z0h values at the TP sites are also larger than at Dunhuang and Audubon during the daytime, the improvements in the revised Noah are believed to be particularly meaningful at two high-elevation sites, as shown in Tables 6 and 7.
6. Conclusions
In this paper we investigated the importance of surface flux parameterization in simulating land surface processes in arid regions. Six thermal surface roughness length schemes were intercompared in the Noah LSM and then the revised version, with one promising z0h scheme evaluated at four arid or semiarid sites. Our major findings are as follows.
The parameterization of z0h is crucial for modeling Tsfc and the surface energy budget in arid regions. Sensitivity tests for the six selected schemes confirm foregoing studies that found the daytime Tsfc is sensitive to the parameterization. If z0h is overestimated (underestimated), Tsfc would be underestimated (overestimated). In addition, H would be overestimated (underestimated), while G0 would be underestimated (overestimated).
The Noah LSM, originally using Z95 as z0h its parameterization scheme, produces unseasonably high z0h during the day in arid regions, which leads to overestimated H and underestimated Tsfc in the daytime. By implementing the Y08 z0h scheme, the revised Noah model can well simulate Tsfc and the surface energy budget simultaneously. Given the wide usage of the Noah LSM in NCEP numerical models, further efforts are warranted to examine potential improvements in synoptic or climatic simulations adopting the presently updated land surface parameterization.
Acknowledgments
We are grateful to three reviewers whose comments have helped the authors to improve the quality of the paper. This work was supported by the National Basic Research Program of China (2009CB421405), Chinese Academy of Sciences (Innovation Project KZCX2-YW-Q11-01 and the “100-Talent” Program), and the National Natural Science Foundation of China (40875009). Audubon Research Ranch data were obtained through the AmeriFlux network. The Shiquanhe and Gaize data used in this paper were obtained through the GAME/Tibet project, which was supported by MEXT, FRSGC, NASDA of Japan, the Chinese Academy of Science, and the Asian Pacific Network. The Dunhuang data were obtained through the Field Experiment on Interaction between Land and Atmosphere in Arid Region of Northwest China. XG gratefully acknowledges the support of the K. C. Wong Education Foundation, Hong Kong.
REFERENCES
Betts, A. K., Chen F. , Mitchell K. E. , and Janjić Z. , 1997: Assessment of the land surface and boundary layer models in the two operational versions of the NCEP Eta Model using FIFE data. Mon. Wea. Rev., 125 , 2896–2916.
Brutsaert, W. H., 1982: Evaporation Into the Atmosphere: Theory, History, and Applications. D. Reidel, 299 pp.
Chen, F., and Dudhia J. , 2001: Coupling an advanced land surface–hydrology model with the PSU–NCAR MM5 modeling system. Part I: Model description and implementation. Mon. Wea. Rev., 129 , 569–585.
Chen, F., and Zhang Y. , 2009: On the coupling strength between the land surface and the atmosphere: From viewpoint of surface exchange coefficients. Geophys. Res. Lett., 36 , L10404. doi:10.1029/2009GL037980.
Chen, F., and Coauthors, 1996: Modeling of land-surface evaporation by four schemes and comparison with FIFE observations. J. Geophys. Res., 101 , 7251–7268.
Chen, F., Janjić Z. , and Mitchell K. E. , 1997: Impact of atmospheric surface-layer parameterizations in the new land-surface scheme of the NCEP mesoscale Eta model. Bound.-Layer Meteor., 85 , 391–421.
Ek, M. B., Mitchell K. E. , Lin Y. , Rogers E. , Grunmann P. , Koren V. , Gayno G. , and Tarpley J. D. , 2003: Implementation of Noah LSM advances in the National Centers for Environmental Prediction operational mesoscale Eta model. J. Geophys. Res., 108 , 8851. doi:10.1029/2002JD003296.
Hogue, T. S., Bastidas L. , Gupta H. , Sorooshian S. , Mitchell K. , and Emmerich W. , 2005: Evaluation and transferability of the Noah LSM in semiarid environments. J. Hydrometeor., 6 , 68–83.
Kanda, M., Kanega M. , Kawai T. , Moriwaki R. , and Sugawara H. , 2007: Roughness lengths for momentum and heat derived from outdoor urban scale models. J. Appl. Meteor. Climatol., 46 , 1067–1079.
Koike, T., Yasunari, Wang J. , and Yao T. , 1999: GAME-Tibet IOP summary report. Proc. First Int. Workshop on GAME–Tibet, Xi’an, China, Chinese Academy of Sciences and Japanese National Committee for GAME, 1–2.
Koren, V., Schaake J. C. , Mitchell K. E. , Duan Q. Y. , Chen F. , and Baker J. , 1999: A parameterization of snowpack and frozen ground intended for NCEP weather and climate models. J. Geophys. Res., 104 , 19569–19585.
LeMone, M. A., Tewari M. , Chen F. , Alfieri J. , and Niyogi D. , 2008: Evaluation of the Noah land surface model using data from a fair-weather IHOP_200 day with heterogeneous surface fluxes. Mon. Wea. Rev., 136 , 4915–4941.
Ma, Y-M., Tsukamoto O. , Wang J. , Ishikawa H. , and Tamagawa I. , 2002: Analysis of aerodynamic and thermodynamic parameters over the grassy marshland surface of Tibetan Plateau. Prog. Natl. Sci., 12 , 36–40.
Ma, Z., and Fu C. , 2006: Some evidences of drying trend over North China from 1951 to 2004. Chin. Sci. Bull., 51 , 2913–2925.
Mahrt, L., 1996: The bulk aerodynamic formulation over heterogeneous surfaces. Bound.-Layer Meteor., 78 , 87–119.
Mahrt, L., and Ek M. , 1984: The influence of atmospheric stability on potential evaporation. J. Climate Appl. Meteor., 23 , 222–234.
Mahrt, L., and Pan H-L. , 1984: A two-layer model of soil hydrology. Bound.-Layer Meteor., 29 , 1–20.
Puigdefabregas, J., 1995: Desertification: Stress beyond resilience, exploring a unifying process structure. Ambio, 24 , 311–313.
Schaake, J. C., Koren V. I. , Duan Q-Y. , Mitchell K. E. , and Chen F. , 1996: Simple water balance model for estimating runoff at different spatial and temporal scales. J. Geophys. Res., 101 , 7461–7475.
Sheppard, P. A., 1958: Transfer across the earth’s surface and through the air above. Quart. J. Roy. Meteor. Soc., 84 , 205–224.
Smeets, C. J. P. P., and van den Broeke M. R. , 2008: The parameterisation of scalar transfer over rough ice. Bound.-Layer Meteor., 128 , 339–355.
Sun, J., 1999: Diurnal variations of thermal roughness height over a grassland. Bound.-Layer Meteor., 92 , 407–427.
Verhoef, A., de Bruin H. A. R. , and van den Hurk B. J. J. M. , 1997: Some practical notes on the parameter kB−1 for sparse vegetation. J. Appl. Meteor., 36 , 560–572.
Warren, A., 1996: Desertification. The Physical Geography of Africa, W. M. Adams, A. S. Goudie, and A. R. Orme, Eds., Oxford University Press, 342–355.
Yang, K., Koike T. , Fujii H. , Tamagawa K. , and Hirose N. , 2002: Improvement of surface flux parameterizations with a turbulence-related length. Quart. J. Roy. Meteor. Soc., 128 , 2073–2087.
Yang, K., Koike T. , and Yang D. , 2003: Surface flux parameterization in the Tibetan Plateau. Bound.-Layer Meteor., 106 , 245–262.
Yang, K., and Coauthors, 2007: Initial CEOP-based review of the prediction skill of operational general circulation models and land surface models. J. Meteor. Soc. Japan, 85 , 99–116.
Yang, K., and Coauthors, 2008: Turbulent flux transfer over bare soil surfaces: Characteristics and parameterization. J. Appl. Meteor. Climatol., 47 , 276–290.
Yang, K., Chen Y-Y. , and Qin J. , 2009: Some practical notes on the land surface modeling in the Tibetan Plateau. Hydrol. Earth Syst. Sci., 13 , 687–701.
Zeng, X., and Dickinson R. E. , 1998: Effect of surface sublayer on surface skin temperature and fluxes. J. Climate, 11 , 537–550.
Zhang, Q., and Coauthors, 2005: NWC-ALIEX and its research advances (in Chinese). Adv. Earth Sci., 20 , 427–441.
Zilitinkevich, S. S., 1995: Non-local turbulent transport: Pollution dispersion aspects of coherent structure of convective flows. Air Pollution Theory and Simulation, H. Power, N. Moussiopoulos, and C. A. Brebbia, Eds., Air Pollution III, Vol. I, Computational Mechanics Publications, 53–60.
Comparison of the diurnal variation of the (a) surface exchange coefficient, (b) sensible heat flux, (c) soil heat flux, and (d) surface temperature among simulations using different z0h schemes: 1) Y08, 2) S58, 3) B82, 4) Z95, 5) Z98, and 6) K07 at Shiquanhe. Note that the observed Tsfc are available in (d).
Citation: Journal of Hydrometeorology 11, 4; 10.1175/2010JHM1185.1
Comparison of the diurnal variation of the (a) surface temperature, (b) net radiation, (c) sensible heat flux, and (d) soil heat flux between two simulations by the revised Noah LSM and the original version against the observations for the Audubon site. Circles represent the observations, the dark gray line represents the simulations by revised the Noah model, and dashed black line represents the simulations by the original model.
Citation: Journal of Hydrometeorology 11, 4; 10.1175/2010JHM1185.1
As in Fig. 2, but for Dunhuang.
Citation: Journal of Hydrometeorology 11, 4; 10.1175/2010JHM1185.1
Comparison of the diurnal variation of the (a) surface temperature, (b) net radiation, (c) sensible heat flux, and (d) soil heat flux from two simulations by the revised Noah and the original version for Shiquanhe.
Citation: Journal of Hydrometeorology 11, 4; 10.1175/2010JHM1185.1
As in Fig. 4, but for Gaize.
Citation: Journal of Hydrometeorology 11, 4; 10.1175/2010JHM1185.1
Simulated mean diurnal variation of ln(z0h) throughout the simulation period at (a) Audubon, (b) Dunhuang, (c) Shiquanhe, and (d) Gaize.
Citation: Journal of Hydrometeorology 11, 4; 10.1175/2010JHM1185.1
General information about the four sites used in this study.
Model parameters derived from observations at four sites. Note that a D in parentheses means a default value.
The z0h parameterization schemes selected for our sensitivity study: Re* = z0mu*/ν, Pr = 0.71, k = 0.4, ν is the fluid kinematical viscosity, α = 0.52, and C = 0.075 in Z95.
Determination coefficient (R2), bias (BIAS), mean deviation (MD), and root-mean-square deviation (RMSD) between the observed Tsfc and the simulations using the schemes in Table 3 for Shiquanhe.
Sensitivity test of the simulated surface temperature (Tsfc) and sensible heat flux (H) to the aerodynamic roughness length (z0m). Comparative statistics are calculated between z0m = 0.011 m and z0m = 0.001 m for Shiquanhe.
The error metrics of the difference between the half-hourly observations and the simulated results at four sites from the revised Noah model and the original version, respectively: BIAS, MD, and RMSD. The observed sensible heat flux was not available at Shiquanhe and Gaize.
