1. Introduction
Summer precipitation over the Tibetan Plateau (TP) and its propagation can result in extreme rainfall and flooding events over the TP and its adjacent areas. Genesis and propagation of summer precipitation over the TP are commonly accompanied by a meso-α-scale boundary layer vortex—the Tibetan Plateau vortex (TPV)—which is closely related to surface diabatic heating effects and boundary layer processes (Ye and Gao 1979; Tao and Ding 1981; Li and Xu 2005; Curio et al. 2019). Better understanding of the processes that influence precipitation evolution from the perspective of land–atmosphere interactions could contribute to better precipitation forecasts over the TP and its vicinity.
Soil moisture affects the partitioning of surface enthalpy fluxes into sensible and latent heat, which influences boundary layer (BL) growth and moisture availability and can have either positive or negative effect on the formation and evolution of convection and precipitation (e.g., Giorgi et al. 1996; Eltahir 1998; Findell and Eltahir 2003a). Studies over the TP have shown that, for instance, surface sensible heating is conducive to local intensification of vortices over the TP (Shen et al. 1986). Different soil moisture states before and after monsoon onset produce different thermodynamic conditions that affect the intensity and development of precipitation over the TP (Yamada and Uyeda 2006). Numerical simulation in Nam Co Lake basin over the TP showed a positive correlation between soil moisture, surface latent heat, and convective development, convective development was strongest with intermediate soil moisture (Gerken et al. 2015). Genesis and eastward propagation of a TPV is usually accompanied by the activity and eastward propagation of horizontal low-level wind shear, which is closely related to surface diabatic heating effects (Li et al. 2017). A TPV cannot form when both surface sensible and latent heat fluxes are removed (Wu et al. 2018). Strong daytime surface diabatic heating, which is essential to the development of daytime convection, provides a favorable condition for nighttime TPV genesis due to the merging of daytime-generated convection near sunset (Zhang et al. 2019a). In summer, soil moisture is heterogeneous over the TP, it is abundant over the southeastern TP and decreases gradually toward the northwestern TP, and the enhanced southeast–northwest gradient of soil moisture distribution over the TP could effectively form convection over the eastern TP (Sugimoto and Ueno 2010). Although responses of precipitation to soil moisture or surface diabatic heating over the TP have been noted, it is not clear how this response works in detail in the intermediate link—that is, the boundary layer.
As a key process in the BL, vertical diffusion, which significantly affects the vertical mixing and directly influences how surface flux transports/redistributes in the BL, is important in the simulation of weather and climate. For instance, discrepancies in BL vertical diffusion can lead to significant simulation biases of general circulation (McGrath-Spangler et al. 2015). Excessive BL vertical diffusion not only forms positive feedback between surface latent heating and convection, changes BL vertical structures accordingly, but also overestimates the Asian summer monsoon’s intensity and precipitation (Cha et al. 2008). Weaker BL vertical diffusion in a global climate model will cause wet biases in the lower atmosphere (Holtslag and Boville 1993). The increased initial soil moisture before a tropical storm’s landfall produces a weaker precipitation after landfall because of the strengthened vertical diffusion within the storm boundary layer over land (Zhang et al. 2019b). Different soil moisture states have a direct influence on the diurnal evolution of BL vertical diffusion outside the storm over land; stronger BL vertical diffusion during the daytime contributes to the maintenance of the precipitation’s symmetric structures after landfall (Zhang et al. 2021).
Above studies have highlighted that BL vertical diffusion could be a key factor that influences the consensus about the impact mechanisms of soil moisture on precipitation. In fact, these mechanisms are strongly influenced by BL’s structures (Findell and Eltahir 2003b). Impacts of BL vertical diffusion are complex and controversial. One crucial issue is that in current BL scheme, vertical eddy diffusivity for heat (Kh) and vapor (Kq), which portrays vertical diffusion processes of heat and moisture, respectively, are generally assumed to be equal, without strictly distinguish (Mellor and Yamada 1974; Nakanishi and Niino 2006; Hong et al. 2006). However, studies based on observation and large-eddy simulation showed that there are obvious differences between Kh and Kq, and there are also different results on which is larger or smaller for Kh and Kq (Laubach et al. 2000; De Roode 2007). Over the TP in summer, potential temperature is well-mixed while moisture is not well-mixed (Yanai and Li 1994), and there are considerable simulation biases of moisture in the lower atmosphere (Bao and Zhang 2013). These imply that the response of boundary layer vertical heat and moisture and precipitation to soil moisture over the TP might be related to Kh and Kq. Ours concerns are 1) how do dry or wet land surface and their spatial heterogeneity over the TP affect precipitation? 2) What are the respective effects of Kh and Kq? Answering these questions would benefit to better understand the response of precipitation to soil moisture over the TP.
The next section introduces the model, data, and experiment design, as well as an overview of three precipitation events. Section 3 compares the results of soil moisture ensemble experiments against observations. Impact mechanisms of initial soil moisture anomaly on precipitation simulation in view of BL vertical diffusion are discussed in section 4. The respective effects of Kh and Kq are further discussed in section 5. Concluding remarks will be made in the final section.
2. Descriptions of the model, data, experiment design, and cases
a. Model, data, and experiment design
An advanced research version of the WRF (ARW) Model (Skamarock et al. 2008) version 3.9.1 is employed for the numerical simulations. Two domains in a one-way nested procedure are used (not shown), and the innermost domain covers the main area of the TP. The horizontal grid spacing of the innermost domain (Fig. 1) is 5 km. A terrain-following η vertical coordinate is adopted in this study, including 41 vertical levels, with the lowest η level at about 27 m above ground level (AGL) over the TP. Based on previous evaluation results for summer precipitation over the TP (e.g., Gao et al. 2015; Wang et al. 2019; Zhang et al. 2019a), physics parameterization schemes selected in this study are: the Kain–Fritsch cumulus scheme (Kain 2004; for the “d01” domain only), the WSM6 microphysics scheme (Hong et al. 2004), the RRTMG scheme (Iacono et al. 2008) for longwave and shortwave radiation processes, the Noah land surface scheme (Chen and Dudhia 2001), and the YSU boundary layer scheme (Hong et al. 2006). The top-down mixing parameterization in the YSU scheme, which is motivated to handle the fog and low-stratus problems (Wilson and Fovell 2018), is deactivated in the simulation. Figure 1 shows the land cover over the TP derived from a MODIS satellite (also used to drive the WRF Model). Results show that four main land-cover categories are present on the TP, that are, barren or sparsely vegetated (24%, referred to “Barren land”), open shrubland (16%, “Shrub land”), grassland (31%, “Grass land”), and multitype forest (7%, “Forest land”) from the northeastern to southeastern TP, respectively. To examine the spatially heterogeneous effects of initial soil moisture perturbations, results with these four land-cover types will be focused below, because these four land-cover types can well reflect the gradual decease in soil moisture from the southeastern to northwestern TP.
Spatial distribution of land-cover types over the TP. The black lines represent the border of different land-cover types over the TP (hereafter). The black dot and triangles denote the in situ observations from TIPEX-III and the in situ conventional soundings (hereafter). The text denotes sounding site names and serial numbers.
Citation: Monthly Weather Review 152, 2; 10.1175/MWR-D-23-0025.1
Initial and boundary conditions are derived from analyses produced by the National Centers for Environmental Prediction (NCEP) FNL. To guarantee enough model spinup for soil moisture, simulations for the three cases are from 0000 UTC 1 August to 3 August 2014, 0000 UTC 15 August to 17 August 2014, and 0000 UTC 18 August to 20 August 2014. The results of first 24 h are not analyzed. Observations used in this study include atmospheric soundings, cloud-top brightness temperature (TBB) from the FY-2E satellite, and GPM precipitation (Huffman et al. 2019) with 0.1° horizontal resolution. Sounding at “Gerze” station (black dot in Fig. 1) comes from the Third Tibetan Plateau Atmospheric Scientific Experiment (TIPEX-III), which has records at 0000, 0600, and 1200 UTC. Other soundings (black triangles in Fig. 1) come from the China Meteorological Administration (CMA), which are conventionally launched twice a day at 0000 and 1200 UTC, additional and intensive soundings at 0600 UTC over the TP are very sparse (e.g., Guo et al. 2016; Zhang et al. 2018), and are missing in the study period (personal communication with Prof. Jianping Guo). Simulations and observations are interpolated at the same grid resolution, data outside the TP are excluded in the whole analysis.
Because of the lack of soil moisture observations in different location of the TP, it is impossible to modify initial soil moisture based on observations. To avoid the influence of soil moisture uncertainty, we shall emphasize mainly on the ensemble mean results. In particular, eleven experiments are designed, that are, “reference run” (referred to REF), which is performed with volumetric water content (VWC) downscaled from the FNL data at initial integration time directly; and “sensitivity runs,” which perturb the initial VWC in steps of 10% ranging from −50% to +50% with respect to REF. In other words, “sensitivity runs” include 10 experiments in total, referred to “−50%, −40%, −30%, −20%, −10%, +10%, +20%, +30%, +40%, and +50%,” respectively. Changes in initial soil moisture are applied to the entire innermost domain and to all soil levels, this keeps any mesoscale disturbances that might develop in the tight soil moisture gradient belts outside of the area of interest (Barthlott and Kalthoff 2011). The ensemble means of the decreasing and increasing initial soil moisture are referred to “ENS_DRY” and “ENS_WET,” respectively.
b. Overview of the three precipitation episodes
In general, the activity of TPV is accompanied by obvious activity of convections, with cloud-top brightness temperature (TBB) less than 220 K and a continuous area larger than 4000 km2 (Yaodong et al. 2008; Sugimoto and Ueno 2010), therefore, TPV can be identified by convections. Figure 2 shows the evolution of TBB of the three precipitation episodes. Results indicate that convections develop rapidly and become vigorous in the north-central TP after sunrise (0600 UTC). From 1000 to 1200 UTC (daytime), a TPV forms and strengthens gradually, with decreased TBB less than 220 K and a continuous area larger than 4000 km2. The TPV decays evidently in the nighttime, with weakened convection. During the weakening of TPV, convections propagate to the southeastern TP. This can also be seen from the in situ observed wind directions at 500 hPa, where westerly wind over the TP increases from 1200 to 2400 UTC. Overall, these cases are related to the activity of TPVs, and their precipitations are located in the central TP in the daytime, while propagating to the southeastern TP in the nighttime. As discussed by Sugimoto and Ueno (2010) and Li et al. (2017), summer precipitations over the TP are concentrated mainly in the southeastern TP and are commonly characterized by the eastward propagation of a thermally induced cyclonic circulation formed in the north-central TP. Therefore, these cases are representative.
The observed cloud-top brightness temperature (unit: K) for precipitation case 1 at (a) 0600 UTC 2 Aug, (b) 1000 UTC 2 Aug, (c) 1200 UTC 2 Aug, (d) 0000 UTC 3 Aug; for precipitation case 2 at (e) 0600 UTC 16 Aug, (f) 1000 UTC 16 Aug, (g) 1200 UTC 16 Aug, (h) 0000 UTC 17 Aug; and for precipitation case 3 at (i) 0600 UTC 19 Aug, (j) 1000 UTC 19 Aug, (k) 1200 UTC 19 Aug, and (l) 0000 UTC 20 Aug. The black vectors denote the wind directions at 500 hPa of in situ stations shown in Fig. 1.
Citation: Monthly Weather Review 152, 2; 10.1175/MWR-D-23-0025.1
3. Comparison of precipitation simulation
Figures 3 and 4 compare the spatial distribution of simulated daytime and nighttime precipitation against observations. Similar to Fig. 2, the observed precipitation in the daytime is located mainly in the central TP (Figs. 3a,e,i); during the nighttime, the observed heavy rainfall exists mainly in the southeastern TP (Figs. 4a,e,i). Simulations generally capture the observed precipitation features in the daytime, and the propagation of precipitation from the central TP to the southeastern TP from daytime to nighttime. Compared to REF, increasing (decreasing) soil moisture produces more (less) rainfall over the TP. Comparisons of precipitation intensity for the three cases (Figs. 5a,c,e) show that all simulations overestimate the observed precipitation. When soil moisture is increased (decreased), the simulated precipitation shifts to higher (lower) values; increasing initial soil moisture produces more precipitation over the TP compared to REF; even for very dry land surfaces, precipitation is still overestimated. For all three cases, both convections and precipitation increase when initial soil moisture is increased, suggesting that a wetter land surface is conducive to more convections and stronger accumulated precipitation over the TP (Figs. 5b,d,f). Similar results can be obtained when applied to different land-cover types (not shown).
Accumulated precipitation (unit: mm) in the daytime (0900–1500 UTC) in (a),(e),(i) GPM; (b),(f),(j) REF; (c),(g),(k) ENS_DRY; and (d),(h),(l) ENS_WET for (a)–(d) precipitation case 1, (e)–(h) precipitation case 2, and (i)–(l) precipitation case 3.
Citation: Monthly Weather Review 152, 2; 10.1175/MWR-D-23-0025.1
As in Fig. 3, but for accumulated precipitation (unit: mm) in the nighttime (1800–2400 UTC).
Citation: Monthly Weather Review 152, 2; 10.1175/MWR-D-23-0025.1
Comparison of (a),(c),(e) time series of mean hourly precipitation (unit: mm) over the TP in simulations and observations; and (b),(d),(f) relative deviation (unit: %) of 24-h accumulated precipitation (red line), number of grid points with precipitation (precipitation > 0 mm; blue line), and number of grid points with convection (TBB < 220 K; green line) in sensitivity runs of soil moisture against the reference run. (a),(b) Precipitation case 1; (c),(d) precipitation case 2; and (e),(f) precipitation case 3. The darker the color in (a), (c), and (e), the larger the deviation of sensitivity runs of soil moisture against REF (hereafter).
Citation: Monthly Weather Review 152, 2; 10.1175/MWR-D-23-0025.1
In general, TPV is active in the BL, with maximum positive vorticity generally presents at about 500 hPa. The differences of wind at 500 hPa averaged over the TP between the sensitivity runs and REF, which can represent the differences of boundary layer wind, are compared in Fig. 6, to investigate whether different soil moisture state can influence the precipitation propagation over the TP. Results indicate that wet and dry land surfaces simulate opposite wind direction over the TP, especially after 0800 UTC. Compared to REF, a wetter (drier) land surface tends to simulate a northwestern (southeastern) wind over the TP. Comparison of geopotential height and wind vectors at 500 hPa further indicate that the low-level circulation (TPV) associated with precipitation propagates faster when the land surface is wetter over the TP (Fig. 7). For instance, at 0800 UTC 19 August (Figs. 7a–c), the northwestern TP is dominated by westerly wind, while the southeastern TP is dominated by southerly wind, without clear horizontal wind shear. At 1200 UTC 19 August (Figs. 7d–f), the TPV presents in north-central TP with closed geopotential height (5875-m isoline and 5865-m isoline) and cyclonic wind. At 1800 UTC 19 August (Figs. 7g–i), the closed 5865-m isoline shift and extend to the eastern TP, with the southwesterly strengthens over the southeastern TP. This indicates that the observed variation in wind direction over the TP (Figs. 2i–l) is generally captured by the simulations. Intercomparisons among REF, ENS_DRY, and ENS_WET show that simulation discrepancies are marginal at 0800 UTC 19 August but are obvious after 1200 UTC 19 August. The geopotential height is lower in the wet simulations than in the dry simulations, suggesting that the intensity of low-level circulation (TPV) is stronger when the land surface is wetter over the TP. Moreover, at 1800 UTC 19 August, the closed 5865-m isoline of geopotential height extends to about 101°E in ENS_WET, 99°E in REF, and 97°E in ENS_DRY; in particular, the closed 5865-m isoline in ENS_WET merges with the trough in the northeastern TP. These results indicate that wetter land surface not only produces stronger low-level circulation (TPV), but also tends to simulate mean northwesterly wind over the TP, these could result in stronger precipitation and is conducive to its southeastward propagation. Similar conclusions can be obtained from other two cases.
Comparison of the time series of the difference in mean wind vectors (unit: ×10−1 m s−1) at 500 hPa over the TP in sensitivity runs of soil moisture against REF (sensitivity runs minus REF) for (a) precipitation case 1, (b) precipitation case 2, and (c) precipitation case 3.
Citation: Monthly Weather Review 152, 2; 10.1175/MWR-D-23-0025.1
Geopotential height (unit: m; blue and red contours denote 5875 and 5865 m, respectively) and wind (shaded colors and vectors; unit: m s−1) at 500 hPa in (a),(d),(g) ENS_DRY; (b),(e),(h) REF; and (c),(f),(i) ENS_WET at (a)–(c) 0800 UTC 19 Aug, (d)–(f) 1200 UTC 19 Aug, and (g)–(i) 1800 UTC 19 Aug. Only the results of precipitation case 3 are shown.
Citation: Monthly Weather Review 152, 2; 10.1175/MWR-D-23-0025.1
To summarize, wetter land surface over the TP not only could produce more convections and accumulated precipitation, but also would accelerate the southeastward propagation of precipitation/convection. Precipitation is overestimated even for very dry land surface.
4. Impact mechanism of soil moisture on precipitation in view of boundary layer vertical diffusion
To illustrate how initial soil moisture perturbation affects precipitation intensity and propagation, the simulated vertical heat and moisture structures in lower atmosphere are compared among REF, ENS_DRY minus REF, and ENS_WET minus REF (Fig. 8). Results show that, in the REF simulation, potential temperature (θ) in the northwestern TP (Barren land and Shrub land) is lower than that in the southeastern TP (Grass land and Forest land), suggesting that northwestern TP is colder than southeastern TP. A wetter (drier) land surface tends to cooling (heating) the lower atmosphere with the decreased (increased) θ. Specifically, in the wet (dry) simulations, the cooling (heating) effects in the lower atmosphere are more obvious in the northwestern TP than in the southeastern TP. As a result, thermal contrast between the cold in the northwestern TP and the warm in the southeastern TP is enhanced over wetter land surface but is weakened over drier land surface, leading to faster southeastward propagation of precipitation in the wet simulations than in the dry simulations. Moreover, compared to a drier land surface, a wetter land surface not only simulates more mixing ratio of vapor (qυ) in the lower atmosphere (e.g., below the maximum height of PBLH), but also leads to more mixing ratio of cloud (qc) near the top of BL; more clouds near the top of BL on wetter land surface could moisten the lower/middle atmosphere and prevent the evaporation of convective turrets through the entrainment/invasion of environmental dry air, which would be conducive to convection development, in agreement with the findings of Zehnder et al. (2006), Kirshbaum (2011) and Barthlott and Kalthoff (2011) etc. Similar conclusions can be obtained from other two cases.
Time vs height cross section of the mixing ratio of vapor (qυ; shaded color; unit: g kg−1), potential temperature (θ; black contours; unit: K), and mixing ratio of cloud (qc; red contours; unit: ×10−2 g kg−1) in (a),(d),(g),(j) ENS_DRY minus REF; (b),(e),(h),(k) REF; and (c),(f),(i),(l) ENS_WET minus REF on (a)–(c) Barren land, (d)–(f) Shrub land, (g)–(j) Grass land, and (j)–(l) Forest land. (left to right) The gray line denotes the planetary boundary layer height (PBLH; unit: m) in ENS_DRY, REF, and ENS_WET, respectively. Only the results of precipitation case 3 are shown. The dashed and solid contours denote negative and positive values, respectively.
Citation: Monthly Weather Review 152, 2; 10.1175/MWR-D-23-0025.1
To further illustrate wetter land surface is conducive to convections, similar to Barthlott and Kalthoff (2011), the convection-related variables are analyzed in Fig. 9. Figures 9a–c show that in REF, the lower atmosphere is convective unstable (∂θ/∂z < 0), and the moist static energy is strong during the daytime; wetter (drier) land surface tends to strengthen (weaken) the convective instability and the moist static energy in the BL, indicating that wetter land surface could provide stronger thermal forcing in the lower atmosphere. Wetter land surface also simulates stronger convective available potential energy (CAPE), weaker convective inhibition (CIN), and lower lifting condensation level (LCL) and level of free convection (LFC) than drier land surface (Figs. 9d,e). Analysis of lower tropospheric lifting for the initiation of deep convection (wdiff = wmax − wCIN, where wmax is the maximum upward vertical velocity between the land surface and the LFC, while
Time vs height cross section of the gradient of θe (shaded color; unit: ×10−3 K m−1) and moist static energy (black contours; ×10−3 J kg−1) in (a) ENS_DRY minus REF, (b) REF, and (c) ENS_WET minus REF over the TP. Time series of relative deviation (unit: %) of (d) CAPE and CIN, (e) LCL and LFC, and (f) number of grid points with wdiff > 0 and mixing ratio of rain (qr) for convection (i.e., qr values when TBB < 220 K) over the TP. The solid and dashed lines in (d)–(f) represent ENS_WET and ENS_DRY experiments, respectively. Other information is as in Fig. 8.
Citation: Monthly Weather Review 152, 2; 10.1175/MWR-D-23-0025.1
Figure 10 compares the observed and simulated vertical profiles of θ and qυ at “Gerze” station. Results show that, although simulations exist dry and cold biases in the BL, which could be related to the uncertainties of background data that is used to drive WRF Model (Bao and Zhang 2013) etc., simulations capture the observed evolution characteristics of heat and moisture, that are, from 0600 to 1200 UTC, θ increases while qυ decreases in the lower atmosphere. Similar to Fig. 8, wetter (drier) land surface tends to cooling (heating) lower atmosphere with decreased (increased) θ; compared to drier land surface, wetter land surface simulates larger qυ in the lower atmosphere, and leads to qυ decreases more rapidly in the upper level (e.g., near 1400–2500 m), which means that condensation of moist air is more evident near the top of BL when land surface is wetter. In other words, a wetter land surface tends to moisten BL, contributing to water vapor saturation near the top of BL.
Comparison of vertical profiles of (a),(c) potential temperature (θ; unit: K) and (b),(d) mixing ratio of vapor (qυ; g kg−1) at “Gerze” station (black dot in Fig. 1) between simulations and observation at (a),(b) 0600 and 1200 UTC 2 Aug and at (c),(d) 0600 and 1200 UTC 19 Aug. The solid and dashed lines denote results at 0600 and 1200 UTC, respectively.
Citation: Monthly Weather Review 152, 2; 10.1175/MWR-D-23-0025.1
To understand how Kh and Kq respond to wet and dry land surface and affect heat and moisture structures in the BL, Fig. 11 compares Kh and Kq over the TP among the experiments. Results show that the features of Kh and Kq are similar, they are much stronger during the daytime while become very weak during the nighttime, Kh is larger than Kq. Besides, a wetter (drier) land surface corresponds to a weaker (stronger) Kh and Kq. The responses of Kh and Kq to soil moisture state could be more important during the daytime than during the nighttime. Also, evident discrepancies of Kh and Kq among the experiments are mainly located near the top of BL. Compared to REF in the daytime, drier (wetter) land surface produces stronger (weaker) Kh and Kq, since θ increases with height (
Time vs height cross section of vertical eddy diffusivity for heat (Kh; color shaded; unit: m2 s−1) and vapor (Kq; black contour lines; unit: m2 s−1) over the TP in (a) ENS_DRY minus REF, (b) REF, and (c) ENS_WET minus REF. The thick gray line denotes PBLH (unit: m) in (a) ENS_DRY, (b) REF, and (c) ENS_WET. Only the results of precipitation case 3 are shown.
Citation: Monthly Weather Review 152, 2; 10.1175/MWR-D-23-0025.1
To summarize, wetter (drier) land surface corresponds to weaker (stronger) BL vertical diffusion, which could lead to water vapor is net gain (loss) while potential temperature is net loss (gain) in the BL. The dry or wet land surface would affect precipitation intensity by influencing water vapor amount in the BL, the saturation near the top of BL, thermal forcing and dynamic lifting in the lower atmosphere, while affect precipitation propagation by influencing (potential) temperature in the BL and its contrast between northwestern and southeastern TP.
5. Effects of boundary layer vertical heat and vapor diffusion
In this section, effects of Kh and Kq on boundary layer heat and moisture and precipitation will be discussed.
Figure 12 shows the evolution of observed heat and moisture structures at “Gerze” station. A distinct simulation bias in the BL is that, from 0600 to 1200 UTC, the simulations can well capture the observed increase of θ, but obviously underestimate the observed decrease of qυ, regardless of dry or wet land surface, suggesting that the simulation of qυ in the BL has large uncertainty. Studies based on observations and large-eddy simulation indicated that there are obvious differences between Kh and Kq, but which one is larger or smaller is controversial (Laubach et al. 2000; De Roode 2007). Also, Yanai and Li (1994) found that over the TP in summer, θ is well-mixed while qυ is not well-mixed in the lower atmosphere. These imply that simulation bias of heat and moisture in the BL could be related to Kh and Kq.
Comparison of “mixing line” (defined as θ vs qυ at different height; Stull 1988) at “Gerze” station (black dot in Fig. 1) between simulations and observation at (a) 0600 and 1200 UTC 19 Aug and at (b) 0600 and 1200 UTC 2 Aug. The thick and thin lines/dots denote results at 0600 and 1200 UTC, respectively. The dark and light colors represent the lower and upper parts of the BL, respectively.
Citation: Monthly Weather Review 152, 2; 10.1175/MWR-D-23-0025.1
Since there are no direct observations of Kh and Kq, to investigate influence of Kh and Kq on the simulation of heat and moisture in the BL and precipitation, additional four sensitivity experiments are designed by perturbing α and β in Eq. (3), that are: 1) Khw_Kqw: weaker Kh and Kq simultaneously (α = 0.5 and β = 0.5), 2) Khs_Kqs: stronger Kh and Kq simultaneously (α = 1.5 and β = 1.5), 3) Khw_Kqs: weaker Kh but stronger Kq (α = 0.5 but β = 1.5), and (4) Khs_Kqw: stronger Kh but weaker Kq (α = 1.5 but β = 0.5). Comparison of Kh and Kq among different experiment shows that (Fig. 13 and Figs. S1 and S2 in the online supplemental material), for all soundings at 0600 UTC, change of Kh and Kq against REF in the sensitivity experiment is evident; although perturbing Kh and Kq can change Km by changing Prandtl numbers, Kh and Kq are at least 2 times larger than Km. For some soundings at 1200 UTC, however, Kh and Kq weaken evidently and even become to zero; Kh and Kq become closer to Km; change of Kh and Kq against REF differs to 0600 UTC, for instance, Khs_Kqw has larger Kh and smaller Kq than REF at 0600 UTC, but both Kh and Kq are smaller in Khs_Kqw than that in REF at 1200 UTC. These imply that effects of Kh and Kq would weaken near sunset.
Change (unit: m2 s−1) of averaged Kh and Kq against REF in the sensitivity experiments at all soundings at (a),(c),(e) 0600 and (b),(d),(f) 1200 UTC for (a),(b) precipitation case 1; (c),(d) precipitation case 2; and (e),(f) precipitation case 3. Soundings used in Fig. 14 are marked as star symbols. Different number denotes different sounding site in Fig. 1.
Citation: Monthly Weather Review 152, 2; 10.1175/MWR-D-23-0025.1
To investigate effects of perturbing the relative intensity of Kh and Kq on boundary layer structures, simulations and observations at six soundings are compared (Fig. 14), that are, “Gerze” at 0600 UTC 19 August and 2 August, “Gerze” and “Mangya” at 1200 UTC 19 August, “Golmud” at 1200 UTC 16 August, and “Golmud” at 1200 UTC 2 August. These soundings are selected because at both 0600 and 1200 UTC, Kh and Kq are larger than Km, and change of Kh and Kq against REF is similar. Results show that, within the PBLH, the simulated heat and moisture are sensitive to the relative intensity of Kh and Kq, and could be generally improved by decreasing Kh and Kq simultaneously or increasing Kh but decreasing Kq; responses of the simulated wind to the relative intensity of Kh and Kq do not have a clear feature, this might be related to weaker Km, large-scale or local-scale synoptic/dynamic forcings, or the transient and intermittent nature of wind itself.
Vertical profiles of (left) potential temperature, (center) mixing ratio of vapor, and (right) wind speed for (a)–(c) “Gerze” station at 0600 UTC 19 Aug, (d)–(f) “Gerze” station at 0600 UTC 2 Aug, (g)–(i) “Gerze” station at 1200 UTC 19 Aug, (j)–(l) “Mangya” station at 1200 UTC 19 Aug, (m)–(o) “Golmud” station at 1200 UTC 16 Aug, and (p)–(r) “Golmud” station at 1200 UTC 2 Aug. The black, orange, blue, green, and red text denotes PBLH in REF, Khw_Kqw, Khs_Kqs, Khw_Kqs, and Khs_Kqw experiments, respectively, at different soundings.
Citation: Monthly Weather Review 152, 2; 10.1175/MWR-D-23-0025.1
To further examine effects of perturbing the relative intensity of Kh and Kq, simulation biases in the boundary layer are compared in Fig. 15. For soundings which Kh and Kq are larger than zero at both 0600 and 1200 UTC (Figs. S1 and S2), Figs. 15a and 15b show that compared to REF, the mean simulation bias of heat and moisture could be reduced by 20.4% and 16.5% in Khw_Kqw experiment and 16.9% and 17.4% in Khs_Kqw experiment; the other two experiments, however, have weaker and even worse influence. For all soundings, Figs. 15d and 15e show that, although the mean improvements of heat and moisture are 21.6% and 9.3% in Khw_Kqw experiment and 8.4% and 7.5% in Khs_Kqw experiment, simulation uncertainty becomes larger, which could be related to vertical diffusion in some soundings is zero (see Figs. S1 and S2), and simulations could be largely influenced by large-scale or local-scale synoptic/dynamic forcings. Figures 15c and 15f show that, effects of perturbing the relative intensity of Kh and Kq on the simulation of wind speed is generally weaker than that of heat and moisture, similar to Fig. 14. Above results suggest that decreasing Kh and Kq simultaneously or increasing Kh but decreasing Kq could improve the simulation of heat and moisture in the boundary layer.
Relative change (unit: %) of simulation bias of (a),(d) potential temperature; (b),(e) mixing ratio of vapor; and (c),(f) wind speed within PBLH in sensitivity experiments against REF. Results are averaged for all precipitation cases at (a)–(c) sounding sites “1-6” and (d)–(f) all soundings. The text denotes mean value. Negative value denotes simulation improvement (decreased bias) against REF.
Citation: Monthly Weather Review 152, 2; 10.1175/MWR-D-23-0025.1
Combined with REF experiment (Fig. 8), further analysis averaged over the whole TP (Fig. 16) shows that, in the daytime, for a specific Kq, the weaker (stronger) Kh would weaken (strengthen) the vertical gradient of θ, leading to the BL tends to be more unstable (stable) (Figs. 16a–d). In other words, Kh affects the atmospheric stability in the BL. For a specific Kh, the weaker Kq would inhibit vertical transport of qυ to the top of BL, resulting in qυ tends to concentrate in the lower BL (not well-mixed qυ in the BL); instead, the stronger Kq would promote vertical transport of qυ to the top of BL, resulting in well-mixed qυ in the BL (Figs. 16e–h). In other words, Kq affects moisture amount and its vertical transport in the BL. It is interesting that weaker Kq tends to produce not well-mixed qυ in the BL, which is in accordance with the results of Yanai and Li (1994). Similar to Figs. 14 and 15, responses of wind to the perturbation of relative intensity of Kh and Kq do not have clear features. Figure 16 also shows that near and after sunset (e.g., 1200 UTC), changes of Kh and Kq against REF in the sensitivity experiments become much weaker and close to Km, implying that effects of Kh and Kq would be marginal. Similar conclusions can be obtained from wet/dry land surface and from other two cases (not shown).
Time vs height cross section of the (a)–(d) potential temperature (θ; shaded color; unit: K) and Kh (red contours; unit: m2 s−1), (e)–(h) mixing ratio of vapor (qυ; shaded color; unit: g kg−1) and Kq (red contours; unit: m2 s−1), (i)–(l) wind speed (shaded color; unit: m s−1) and Km (red contours; unit: m2 s−1), (m)–(p) gradient of θe (shaded color; unit: ×10−3 K m−1) and convergence (red contours; unit: 10−5 s−1), and (q)–(t) mixing ratio of cloud (qc; shaded color; unit: ×10−2 g kg−1) over the TP in (a),(e),(i),(m),(q) Khw_Kqw minus REF; (b),(f),(j),(n),(r) Khs_Kqs minus REF; (c),(g),(k),(o),(s) Khw_Kqs minus REF; and (d),(h),(l),(p),(t) Khs_Kqw minus REF. The black line denotes PBLH (unit: m) in each sensitivity experiment of Kh and Kq. Positive and negative contours in (m)–(p) denote weakened and strengthened convergence, respectively. Only the results of precipitation case 3 are shown.
Citation: Monthly Weather Review 152, 2; 10.1175/MWR-D-23-0025.1
Figure 17 investigates impacts of perturbing the relative intensity of Kh and Kq on the precipitation intensity simulation. Results show that Khs_Kqw and Khw_Kqw could effectively alleviate the overestimated precipitation in REF; the other two experiments, however, have marginal influences on precipitation or even tend to aggravate the overestimated precipitation. Responses of convection to the relative intensity of Kh and Kq are basically the same as precipitation (not shown), suggesting that the relative intensity of Kh and Kq has evidently impacts on convective precipitation. These features are consistent to Figs. 14 and 15, suggesting that simulation of precipitation is related to simulation of boundary layer heat and moisture induced by the relative intensity of Kh and Kq. It is found that Kh and Kq have competitive effects on precipitation intensity, that are, precipitation simulation would be improved under specific intensity of Kh and Kq. Weakening Kq could remedy the overestimated precipitation intensity, regardless of Kh values, while Kq is stronger, its influence on precipitation intensity could be related to Kh values, implying that the competitive effects could be manifested by the relative importance/contribution of moisture and atmospheric stability conditions in the BL. For instance, weakening Kh and Kq (Khw_Kqw) simultaneously prevents the vertical transport of water vapor to the top of BL (Fig. 16e); however, BL tends to be more convective unstable because of the stronger heating and moistening in the lower BL (Figs. 16a,m), this is not conducive to precipitation from the perspective of water vapor, but is conducive to precipitation from the perspective of atmospheric stability conditions. Strengthening Kh but weakening Kq would deactivate the competitive effects, that is, compared to Khw_Kqw, Khs_Kqw not only weakens vertical transport of water vapor to the top of BL (Fig. 16h), but also produces less convective unstable BL because of the stronger θ is manifested in the middle and higher BL (Figs. 16d,p), that is, both water vapor and atmospheric stability conditions are not beneficial to precipitation. Accordingly, qc near the top of BL in Khs_Kqw (Fig. 16t) is less than Khw_Kqw (Fig. 16q), which in turn affects convective precipitation by influencing entrainment/invasion of environmental dry air in the lower/middle atmosphere. Adjusting the relative intensity of Kh and Kq also changes the convergence in the lower atmosphere, although the strengthened convergence (Figs. 16m,p) in Khw_Kqw and Khs_Kqw is beneficial to precipitation from the perspective of dynamic lifting, combined with the weakened precipitation in Khw_Kqw and Khs_Kqw (Fig. 17), these suggest that precipitation intensity could be closely related to the competitive effects of Kh and Kq on boundary layer heat and moisture, effects of wind and convergence associated with Km could be relatively weak, which are in accordance with Figs. 14 and 15. Similar conclusions can be analyzed from other experiments.
Relative change (unit: %) of simulation bias of accumulated precipitation from 0600 to 1500 UTC over the TP in sensitivity simulations of K against REF for the three precipitation cases. Negative value denotes simulation improvement (decreased bias) against REF.
Citation: Monthly Weather Review 152, 2; 10.1175/MWR-D-23-0025.1
6. Conclusions
Based on a series of sensitivity experiments, this study investigates the influences of initial soil moisture anomalies on the precipitation simulation over the TP, with the emphasis on the effects of boundary layer vertical diffusion for heat (Kh) and vapor (Kq). The major results are summarized as follows:
Results from three precipitation cases indicate that, wetter land surface over the TP not only could strengthen convections and accumulated precipitation, but also would be conducive to the southeastward propagation of precipitation.
The wetter land surface corresponds to weaker boundary layer vertical diffusion, especially near the top of BL, which not only could promote the net gain of moisture in the BL and encourage air saturation near the top of BL to prevent the environmental dry air entrainment/invasion, but also could strengthen thermal forcing and dynamic lifting in the lower atmosphere, which would be beneficial to more convection and precipitation. The responses of heating to dry or wet land surface are heterogeneous over the TP, wetter land surface could enhance the contrast between the cold in the northwestern TP and the warm in the southeastern TP, which would be conducive to the southeastward propagation of precipitation.
Effects of Kh and Kq on boundary layer heat and moisture are prominent mainly in the daytime. Although adjusting the relative intensity of Kh and Kq could change Km, changes of Km are much smaller than Kh and Kq. The simulation of heat and moisture in the boundary layer could be improved by decreasing Kh and Kq simultaneously or increasing Kh but decreasing Kq. From the perspective of heat and moisture, Kh affects atmospheric stability, while Kq affects moisture amount and its vertical transport in the BL; Kh and Kq have competitive effects on precipitation intensity simulation by influencing the relative importance of moisture and atmospheric stability conditions in the BL. Adjusting the relative intensity of Kh and Kq would deactivate the competitive effects. For instance, weakening Kh and Kq simultaneously is not conducive to precipitation from the perspective of water vapor, but is conducive to precipitation from the perspective of atmospheric stability conditions; instead, strengthening Kh but weakening Kq not only could attenuate convective instability in the BL, but also inhibit vertical transport of moisture to the top of BL and is not beneficial to air saturation there, these would alleviate the overestimated precipitation because both moisture and atmospheric stability conditions are not beneficial to convections.
Although our study suggests that adjusting the relative intensity of Kh and Kq could be important to convection/precipitation simulation over the TP; however, perturbing Kh and Kq changes Prandtl number and Km simultaneously, these could change dynamic processes etc. to influence convection/precipitation simulation, for instance, in Fig. 16, wind and convergence in the BL that are related to Km has discrepancies among the sensitivity experiments of Kh and Kq, implying that the mutual feedback among Kh, Kq, and Km could be important. These issues are beyond the scope of this study and will be discussed in our future work, using more observations (especially at 0600 UTC) and/or conducting large-eddy simulation, to quantify the optimal relation among Kh, Kq, and Km.
Acknowledgments.
This study is supported by the National Science Foundation of China (42275004, 42175064, 91837205), National Key Research and Development Program of China (2023YFC3200041), Natural Science Foundation of Gansu Province of China (20JR5RA309), and the Fundamental Research Funds for the Central Universities (lzujbky-2021-15). Authors are grateful for efforts by the National Center for Atmospheric Research (NCAR) in making the community research version of the WRF Model available on the public website. Comments from two anonymous reviewers were very helpful in improving the manuscript. Computational support by the Supercomputing Center of Lanzhou University is appreciated.
Data availability statement.
FNL data come from https://rda.ucar.edu/; atmospheric sounding at “Gerze” station comes from the Third Tibetan Plateau Atmospheric Scientific Experiment (TIPEX-III, https://data.tpdc.ac.cn/); other soundings and cloud-top brightness temperature come from CMA (http://data.cma.cn/); GPM precipitation data come from https://disc.gsfc.nasa.gov/datasets/GPM_3IMERGHH_06/summary.
REFERENCES
Bao, X., and F. Zhang, 2013: Evaluation of NCEP–CFSR, NCEP–NCAR, ERA-Interim, and ERA-40 reanalysis datasets against independent sounding observations over the Tibetan Plateau. J. Climate, 26, 206–214, https://doi.org/10.1175/JCLI-D-12-00056.1.
Barthlott, C., and N. Kalthoff, 2011: A numerical sensitivity study on the impact of soil moisture on convection-related parameters and convective precipitation over complex terrain. J. Atmos. Sci., 68, 2971–2987, https://doi.org/10.1175/JAS-D-11-027.1.
Cha, D. H., D. K. Lee, and S. Y. Hong, 2008: Impact of boundary layer processes on seasonal simulation of the East Asian summer monsoon using a regional climate model. Meteor. Atmos. Phys., 100, 53–72, https://doi.org/10.1007/s00703-008-0295-6.
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, https://doi.org/10.1175/1520-0493(2001)129<0569:CAALSH>2.0.CO;2.
Curio, J., R. Schiemann, K. Hodges, and A. Turner, 2019: Climatology of Tibetan Plateau vortices in reanalysis data and a high-resolution global climate model. J. Climate, 32, 1933–1950, https://doi.org/10.1175/JCLI-D-18-0021.1.
De Roode, S. R., 2007: The role of eddy diffusivity profiles on stratocumulus liquid water path biases. Mon. Wea. Rev., 135, 2786–2793, https://doi.org/10.1175/MWR3426.1.
Eltahir, E. A., 1998: A soil moisture–rainfall feedback mechanism: 1. Theory and observations. Water Resour. Res., 34, 765–776, https://doi.org/10.1029/97WR03499.
Findell, K. L., and E. A. Eltahir, 2003a: Atmospheric controls on soil moisture–boundary layer interactions. Part I: Framework development. J. Hydrometeor., 4, 552–569, https://doi.org/10.1175/1525-7541(2003)004<0552:ACOSML>2.0.CO;2.
Findell, K. L., and E. A. Eltahir, 2003b: Atmospheric controls on soil moisture–boundary layer interactions: Three-dimensional wind effects. J. Geophys. Res., 108, 8385, https://doi.org/10.1029/2001JD001515.
Gao, Y., J. Xu, and D. Chen, 2015: Evaluation of WRF mesoscale climate simulations over the Tibetan Plateau during 1979–2011. J. Climate, 28, 2823–2841, https://doi.org/10.1175/JCLI-D-14-00300.1.
Gerken, T., W. Babel, M. Herzog, K. Fuchs, F. Sun, Y. Ma, T. Foken, and H. F. Graf, 2015: High-resolution modelling of interactions between soil moisture and convective development in a mountain enclosed Tibetan Basin. Hydrol. Earth Syst. Sci., 19, 4023–4040, https://doi.org/10.5194/hess-19-4023-2015.
Giorgi, F., L. O. Mearns, C. Shields, and L. Mayer, 1996: A regional model study of the importance of local versus remote controls of the 1988 drought and 1993 flood over the central United States. J. Climate, 9, 1150–1162, https://doi.org/10.1175/1520-0442(1996)009<1150:ARMSOT>2.0.CO;2.
Guo, J., and Coauthors, 2016: The climatology of planetary boundary layer height in China derived from radiosonde and reanalysis data. Atmos. Chem. Phys., 16, 13 309–13 319, https://doi.org/10.5194/acp-16-13309-2016.
Holtslag, A. A. M., and B. A. Boville, 1993: Local versus nonlocal boundary-layer diffusion in a global climate model. J. Climate, 6, 1825–1842, https://doi.org/10.1175/1520-0442(1993)006<1825:LVNBLD>2.0.CO;2.
Hong, S. Y., J. Dudhia, and S. H. Chen, 2004: A revised approach to ice microphysical processes for the bulk parameterization of clouds and precipitation. Mon. Wea. Rev., 132, 103–120, https://doi.org/10.1175/1520-0493(2004)132<0103:ARATIM>2.0.CO;2.
Hong, S. Y., Y. Noh, and J. Dudhia, 2006: A new vertical diffusion package with an explicit treatment of entrainment processes. Mon. Wea. Rev., 134, 2318–2341, https://doi.org/10.1175/MWR3199.1.
Huffman, G. J., E. F. Stocker, D. T. Bolvin, E. J. Nelkin, and J. Tan, 2019: GPM IMERG final precipitation L3 half hourly 0.1 degree × 0.1 degree V06. Goddard Earth Sciences Data and Information Services Center (GES DISC), accessed 25 October 2022, https://doi.org/10.5067/GPM/IMERG/3B-HH/06.
Iacono, M. J., J. S. Delamere, E. J. Mlawer, M. W. Shephard, S. A. Clough, and W. D. Collins, 2008: Radiative forcing by long-lived greenhouse gases: Calculations with the AER radiative transfer models. J. Geophys. Res., 113, D13103, https://doi.org/10.1029/2008JD009944.
Kain, J. S., 2004: The Kain–Fritsch convective parameterization: An update. J. Appl. Meteor., 43, 170–181, https://doi.org/10.1175/1520-0450(2004)043<0170:TKCPAU>2.0.CO;2.
Kirshbaum, D. J., 2011: Cloud-resolving simulations of deep convection over a heated mountain. J. Atmos. Sci., 68, 361–378, https://doi.org/10.1175/2010JAS3642.1.
Laubach, J., K. G. McNaughton, and J. D. Wilson, 2000: Heat and water vapour diffusivities near the base of a disturbed stable internal boundary layer. Bound.-Layer Meteor., 94, 23–63, https://doi.org/10.1023/A:1002402515319.
Li, G., and Q. Xu, 2005: Effect of dynamic pumping in the boundary layer on the Tibetan Plateau vortices. Chin. J. Atmos. Sci., 29, 965–972, https://doi.org/10.3878/j.issn.1006-9895.2005.06.12.
Li, G., S. Li, and C. Huang, 2017: Research status and prospect of the interaction between Tibetan Plateau shear line and Tibetan Plateau vortex. Diqiu Kexue Jinzhan, 32, 919–925.
McGrath-Spangler, E. L., A. Molod, L. E. Ott, and S. Pawson, 2015: Impact of planetary boundary layer turbulence on model climate and tracer transport. Atmos. Chem. Phys., 15, 7269–7286, https://doi.org/10.5194/acp-15-7269-2015.
Mellor, G. L., and T. Yamada, 1974: A hierarchy of turbulence closure models for planetary boundary layers. J. Atmos. Sci., 31, 1791–1806, https://doi.org/10.1175/1520-0469(1974)031<1791:AHOTCM>2.0.CO;2.
Nakanishi, M., and H. Niino, 2006: An improved Mellor–Yamada level 3 model: Its numerical stability and application to a regional prediction of advecting fog. Bound.-Layer Meteor., 119, 397–407, https://doi.org/10.1007/s10546-005-9030-8.
Shen, R., E. R. Reiter, and J. F. Bresch, 1986: Some aspects of the effects of sensible heating on the development of summer weather systems over the Tibetan Plateau. J. Atmos. Sci., 43, 2241–2260, https://doi.org/10.1175/1520-0469(1986)043<2241:SAOTEO>2.0.CO;2.
Skamarock, W. C., and Coauthors, 2008: A description of the Advanced Research WRF version 3. NCAR Tech. Note NCAR/TN-475+STR, 113 pp., https://doi.org/10.5065/D68S4MVH.
Stull, R. B., 1988: An Introduction to Boundary Layer Meteorology. Kluwer Academic, 666 pp.
Sugimoto, S., and K. Ueno, 2010: Formation of mesoscale convective systems over the eastern Tibetan Plateau affected by plateau-scale heating contrasts. J. Geophys. Res., 115, D16105, https://doi.org/10.1029/2009JD013609.
Tao, S. Y., and Y. H. Ding, 1981: Observational evidence of the influence of the Qinghai–Xizang (Tibet) Plateau on the occurrence of heavy rain and severe convective storms in China. Bull. Amer. Meteor. Soc., 62, 23–30, https://doi.org/10.1175/1520-0477(1981)062<0023:OEOTIO>2.0.CO;2.
Wang, C., D. Wu, and F. Zhang, 2019: Modification of the convective adjustment time scale in the Kain–Fritsch eta scheme for the case of weakly forced deep convection over the Tibetan Plateau region. Quart. J. Roy. Meteor. Soc., 145, 1915–1932, https://doi.org/10.1002/qj.3535.
Wilson, T. H., and R. G. Fovell, 2018: Modeling the evolution and life cycle of radiative cold pools and fog. Wea. Forecasting, 33, 203–220, https://doi.org/10.1175/WAF-D-17-0109.1.
Wu, D., F. Zhang, and C. Wang, 2018: Impacts of diabatic heating on the genesis and development of an inner Tibetan Plateau vortex. J. Geophys. Res. Atmos., 123, 11 691–11 704, https://doi.org/10.1029/2018JD029240.
Yamada, H., and H. Uyeda, 2006: Transition of the rainfall characteristics related to the moistening of the land surface over the Central Tibetan Plateau during the summer of 1998. Mon. Wea. Rev., 134, 3230–3247, https://doi.org/10.1175/MWR3235.1.
Yanai, M., and C. Li, 1994: Mechanism of heating and the boundary layer over the Tibetan Plateau. Mon. Wea. Rev., 122, 305–323, https://doi.org/10.1175/1520-0493(1994)122<0305:MOHATB>2.0.CO;2.
Yaodong, L., W. Yun, S. Yang, H. Liang, G. Shouting, and R. Fu, 2008: Characteristics of summer convective systems initiated over the Tibetan Plateau. Part I: Origin, track, development, and precipitation. J. Appl. Meteor. Climatol., 14, 2679–2695, https://doi.org/10.1175/2008JAMC1695.1.
Ye, D. Z., and Y. X. Gao, 1979: The Tibetan Plateau Meteorology (in Chinese). Science Press, 278 pp.
Zehnder, J. A., L. Zhang, D. Hansford, A. Radzan, N. Selover, and C. M. Brown, 2006: Using digital cloud photogrammetry to characterize the onset and transition from shallow to deep convection over orography. Mon. Wea. Rev., 134, 2527–2546, https://doi.org/10.1175/MWR3194.1.
Zhang, F., C. Wang, and Z. Pu, 2019a: Genesis of Tibetan Plateau vortex: Roles of surface diabatic and atmospheric condensational latent heating. J. Appl. Meteor. Climatol., 58, 2633–2651, https://doi.org/10.1175/JAMC-D-19-0103.1.
Zhang, F., Z. Pu, and C. Wang, 2019b: Impacts of soil moisture on the numerical simulation of a post-landfall storm. J. Meteor. Res., 33, 206–218, https://doi.org/10.1007/s13351-019-8002-8.
Zhang, F., Z. Pu, and C. Wang, 2021: Land-surface diurnal effects on the asymmetric structures of a post-landfall tropical storm. J. Geophys. Res Atmos., 126, 2020JD033842, https://doi.org/10.1029/2020JD033842.
Zhang, W., and Coauthors, 2018: On the summertime planetary boundary layer with different thermodynamic stability in China: A radiosonde perspective. J. Climate, 31, 1451–1465, https://doi.org/10.1175/JCLI-D-17-0231.1.