• Alpert, P., , and H. Shafir, 1989: Mesoγ-scale distribution of orographic precipitation: Numerical study and comparison with precipitation derived from radar measurements. J. Appl. Meteor., 28 , 11051117.

    • Search Google Scholar
    • Export Citation
  • Barstad, I., , and R. B. Smith, 2005: Evaluation of an orographic precipitation model. J. Hydrometeor., 6 , 8599.

  • Bruintjes, R. T., , T. Clark, , and W. D. Hall, 1994: Interactions between topographic airflow and cloud/precipitation development during the passage of a winter storm in Arizona. J. Atmos. Sci., 51 , 4867.

    • Search Google Scholar
    • Export Citation
  • Colle, B. A., 2004: Sensitivity of orographic precipitation to changing ambient conditions and terrain geometries: An idealized modeling perspective. J. Atmos. Sci., 61 , 588606.

    • Search Google Scholar
    • Export Citation
  • Colle, B. A., , and Y. Zeng, 2004a: Bulk microphysical sensitivities within the MM5 for orographic precipitation. Part I: The Sierra 1986 event. Mon. Wea. Rev., 132 , 27802801.

    • Search Google Scholar
    • Export Citation
  • Colle, B. A., , and Y. Zeng, 2004b: Bulk microphysical sensitivities within the MM5 for orographic precipitation. Part II: Impact of barrier width and freezing level. Mon. Wea. Rev., 132 , 28022815.

    • Search Google Scholar
    • Export Citation
  • Doyle, J. D., and Coauthors, 2000: An intercomparison of model-predicted wave breaking for the 11 January 1972 Boulder windstorm. Mon. Wea. Rev., 128 , 901914.

    • Search Google Scholar
    • Export Citation
  • Durran, D., 2003: Lee waves and mountain waves. Encyclopedia of Atmospheric Sciences, J. R. Holton, J. A. Curry, and J. A. Pyle, Eds., Academic Press, 1161–1170.

    • Search Google Scholar
    • Export Citation
  • Durran, D., , and J. Klemp, 1982: On the effects of moisture on the Brunt-Väisälä frequency. J. Atmos. Sci., 39 , 21522158.

  • Garvert, M. F., , B. A. Colle, , and C. F. Mass, 2005: The 13–14 December 2001 IMPROVE-2 event. Part I: Synoptic and mesoscale evolution and comparison with a mesoscale model simulation. J. Atmos. Sci., 62 , 34743492.

    • Search Google Scholar
    • Export Citation
  • Garvert, M. F., , B. Smull, , and C. Mass, 2007: Multiscale mountain waves influencing a major orographic precipitation event. J. Atmos. Sci., 64 , 711737.

    • Search Google Scholar
    • Export Citation
  • Houze, R., , and S. Medina, 2005: Turbulence as a mechanism for orographic precipitation enhancement. J. Atmos. Sci., 62 , 35993623.

  • Klemp, J. B., , and D. R. Durran, 1983: An upper boundary condition permitting internal gravity wave radiation in numerical mesoscale models. Mon. Wea. Rev., 111 , 430444.

    • Search Google Scholar
    • Export Citation
  • Medina, S., , and R. Houze, 2003: Air motions and precipitation growth in alpine storms. Quart. J. Roy. Meteor. Soc., 129 , 345371.

  • Rasmussen, R. M., , I. Geresdi, , G. Thompson, , K. Manning, , and E. Karplus, 2002: Freezing drizzle formation in stably stratified layer clouds: The role of radiative cooling of cloud droplets, cloud condensation nuclei, and ice initiation. J. Atmos. Sci., 59 , 837860.

    • Search Google Scholar
    • Export Citation
  • Rotunno, R., , and R. Ferretti, 2001: Mechanisms of intense Alpine rainfall. J. Atmos. Sci., 58 , 17321749.

  • Sinclair, M. R., 1994: A diagnostic model for estimating orographic precipitation. J. Appl. Meteor., 33 , 11631175.

  • Smith, R. B., , and I. Barstad, 2004: A linear theory of orographic precipitation. J. Atmos. Sci., 61 , 13771391.

  • Smith, R. B., , Q. Jiang, , M. Fearon, , P. Tabary, , M. Dorninger, , J. Doyle, , and R. Benoit, 2003: Orographic precipitation and air mass transformation: An alpine example. Quart. J. Roy. Meteor. Soc., 129 , 433454.

    • Search Google Scholar
    • Export Citation
  • Smith, R. B., , I. Barstad, , and L. Bonneau, 2005: Orographic precipitation and Oregon’s climate transition. J. Atmos. Sci., 62 , 177191.

    • Search Google Scholar
    • Export Citation
  • Thompson, G., , R. M. Rasmussen, , and K. Manning, 2004: Explicit forecasts of winter precipitation using an improved bulk microphysics scheme. Part I: Description and sensitivity analysis. Mon. Wea. Rev., 132 , 519542.

    • Search Google Scholar
    • Export Citation
  • View in gallery

    (a) Terrain profile (m) used in the idealized simulations for the 0, 1, and 4 windward ridge experiments for U = 15 m s−1. (b) The 6–12-h precipitation across the barrier for the 0, 1, and 4 windward ridge experiments for U = 15 m s−1. The dashed region AB in (a) represents the region where the average precipitation and drying ratio were calculated.

  • View in gallery

    (a), (b) Same as in Fig. 1 but for the 0, 8, 12, and 16 windward ridge experiments.

  • View in gallery

    Cross section for (a) 0, (b) 1, and (c) 4 windward ridges showing snow (gray every 0.04 g kg−1), graupel (dashed every 0.04 g kg−1), and rain (thin solid every 0.04 g kg−1), as well as potential temperature (solid every 4 K) and circulation vectors using the inset scale in (a).

  • View in gallery

    Same as in Fig. 3 but for (a) 8, (b) 12, and (c) 16 windward ridges.

  • View in gallery

    (a) Precipitation (mm) for the 6–12-h period vs ridge number averaged between points A and B in Fig. 1a for the U = 8 m s−1 (U8, dashed), 15 m s−1 (U15, black solid), and 30 m s−1 (U30, gray) experiments. The gray line is the average precipitation for the U15 run using a 2400-m mountain of 85-km half-width and n = 0. The wavelength of the windward ridges is also plotted along the x axis as well as the expected transition to evanescent gravity waves for the U30 (gray), U15 (solid black), and U8 (dashed). (b) Same as in (a) but for the drying ratio.

  • View in gallery

    The 6–12-h precipitation across the barrier for the n = 12 experiment for U = 15 m s−1 (gray lines) and an experiment (N12ALT in Table 1) in which the only every other ridge was used (black lines).

  • View in gallery

    Maximum surface precipitation (mm) accumulated during the 6–12-h period for any point between points A and B in Fig. 1a for the U = 8 m s−1 (U8, dashed), 15 m s−1 (U15, black solid), and 30 m s−1 (U30, gray) experiments. The plotted numbers show the distance (km) upstream from the crest where the maximum precipitation occurred.

  • View in gallery

    Mass-weighted vertical motion (solid, m s−1) averaged in the vertical vs distance for ridge numbers (a) 0, (b)4, (c) 8, and (d) 12 for the U15 (black) and U30 (gray) experiments in Table 1. Terrain is dashed for reference.

  • View in gallery

    Vertically integrated (a) cloud water, (b) rain, (c) snow, and (d) graupel (kg m−2) averaged between points A and B in Fig. 1a. The wavelength of the windward ridges is also plotted along the x axis in (a) as well as the expected transition to evanescent gravity waves for the U30 (gray), U15 (solid black), and U8 (dashed).

  • View in gallery

    Residence time (s) vs ridge number for all hydrometeors between points A and B for the U8 (dashed), U15 (solid black), and U30 (gray) experiments.

  • View in gallery

    (a) Precipitation (mm) for the 6–12-h period vs ridge number averaged between points A and B in Fig. 1a for Nm = 0.01 s−1 and U = 8 m s−1 (N01U8, dashed), 15 m s−1 (N01U15, black solid), and 30 m s−1 (N01U30, gray). (b) Same as in (a) but for the drying ratio.

  • View in gallery

    Cross section for n = 8 windward ridges for the (a) N01U8, (b) N01U15, and (c) N01U30 experiments showing snow (gray every 0.04 g kg−1), graupel (dashed every 0.04 g kg−1), and rain (thin solid every 0.04 g kg−1), as well as potential temperature (solid every 8 K) and circulation vectors using the inset scale.

  • View in gallery

    Fractional precipitation change as compared to the 0 ridge run vs moist Froude number for the 8, 12, and 16 ridge (see numbers on plot) simulations using the N01U8–N01U30 (gray lines) and U8–30 (black line) experiments. The precipitation is averaged between points A and B in Fig. 1a and accumulated for the 6–12-h simulation period.

  • View in gallery

    (a) Precipitation (mm) for the 6–12-h period vs ridge number averaged between points A and B in Fig. 1a for the ridge height experiments of Z = 200 m (Z200, dashed), 400 m (Z400, solid black), and 800 m (Z800, gray) experiments. (b) Same as in (a) but for the drying ratio.

  • View in gallery

    (a) Precipitation (mm) for the 6–12-h period vs ridge number averaged between points A and B in Fig. 1a for the freezing-level experiments of FL= 1000 mb (FL1000, dashed), 750 mb (FL750, solid black), and 500 mb (FL500, gray) experiments. (b) Same as in (a) but for the drying ratio.

  • View in gallery

    The 6–12-h precipitation (mm) across the barrier for the 0, 4, 8, and 16 windward ridge experiments for (a) FL500 and (b) FL1000.

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 6 6 2
PDF Downloads 3 3 1

Two-Dimensional Idealized Simulations of the Impact of Multiple Windward Ridges on Orographic Precipitation

View More View Less
  • 1 School of Marine and Atmospheric Sciences, Stony Brook University, Stony Brook, New York
© Get Permissions
Full access

Abstract

This paper presents two-dimensional (2D) idealized simulations at 1-km grid spacing using the fifth-generation Pennsylvania State University–National Center for Atmospheric Research (PSU–NCAR) Mesoscale Model (MM5) in order to illustrate how a series of ridges along a broad windward slope can impact the precipitation distribution and simulated microphysics. The number of windward ridges for a 2000-m mountain of 50-km half-width is varied from 0 to 16 over a 150-km distance using different stratifications, freezing levels, uniform ambient flows, and ridge amplitudes.

A few (200–400 m) windward ridges can enhance the precipitation locally over each ridge crest by a factor of 2–3. Meanwhile, a series of 8–16 ridges that are 200–400 m in height can increase the net precipitation averaged over the windward slope by 10%–35%. This average precipitation enhancement is maximized when the ridge spacing is relatively small (<20 km), since there is less time for subsidence drying within the valleys and the mountain waves become more evanescent, which favors a simple upward and downward motion couplet over each ridge. In addition, small ridge spacing is shown to have a synergistic effect on precipitation over the lower windward slope, in which an upstream ridge helps increase the precipitation over the adjacent downwind ridge. There is little net precipitation enhancement by the ridges for small moist Froude numbers (Fr < 0.8), since flow blocking limits the flow up and over each ridge. For a series of narrow ridges (∼10 km wide), the largest precipitation enhancement for a 500-mb freezing level occurs over lower windward slope of the barrier through warm-rain processes. In contrast, a 1000-mb freezing level has the largest precipitation enhancement over the middle and upper portions of a barrier for a series of narrow (∼10 km wide) ridges given the horizontal advection of snow aloft.

Corresponding author address: Dr. Brian A. Colle, School of Marine and Atmospheric Sciences, Stony Brook University, Stony Brook, NY 11794-5000. Email: brian.colle@stonybrook.edu

Abstract

This paper presents two-dimensional (2D) idealized simulations at 1-km grid spacing using the fifth-generation Pennsylvania State University–National Center for Atmospheric Research (PSU–NCAR) Mesoscale Model (MM5) in order to illustrate how a series of ridges along a broad windward slope can impact the precipitation distribution and simulated microphysics. The number of windward ridges for a 2000-m mountain of 50-km half-width is varied from 0 to 16 over a 150-km distance using different stratifications, freezing levels, uniform ambient flows, and ridge amplitudes.

A few (200–400 m) windward ridges can enhance the precipitation locally over each ridge crest by a factor of 2–3. Meanwhile, a series of 8–16 ridges that are 200–400 m in height can increase the net precipitation averaged over the windward slope by 10%–35%. This average precipitation enhancement is maximized when the ridge spacing is relatively small (<20 km), since there is less time for subsidence drying within the valleys and the mountain waves become more evanescent, which favors a simple upward and downward motion couplet over each ridge. In addition, small ridge spacing is shown to have a synergistic effect on precipitation over the lower windward slope, in which an upstream ridge helps increase the precipitation over the adjacent downwind ridge. There is little net precipitation enhancement by the ridges for small moist Froude numbers (Fr < 0.8), since flow blocking limits the flow up and over each ridge. For a series of narrow ridges (∼10 km wide), the largest precipitation enhancement for a 500-mb freezing level occurs over lower windward slope of the barrier through warm-rain processes. In contrast, a 1000-mb freezing level has the largest precipitation enhancement over the middle and upper portions of a barrier for a series of narrow (∼10 km wide) ridges given the horizontal advection of snow aloft.

Corresponding author address: Dr. Brian A. Colle, School of Marine and Atmospheric Sciences, Stony Brook University, Stony Brook, NY 11794-5000. Email: brian.colle@stonybrook.edu

1. Introduction

The spatial distribution of orographic precipitation is important for hydrologic forecasting, water resources, regional climate, and ecology. It is well known that the distribution of precipitation in mountainous terrain is highly dependent on the width and height of the terrain (Colle 2004; Colle and Zeng 2004b), since mountain steepness impacts the vertical motion magnitude over the windward slope. As a result, it has been shown that simple upslope-advection models can estimate the condensation rate above the topography by including the terrain slope, wind speed, temperature and moisture profile, and hydrometeor advection (Alpert and Shafir 1989; Sinclair 1994). However, this model assumes no flow blocking and that the vertical motions generated by the mountain are uniform with height. This does not conform to the vertical motion variability introduced by gravity waves over the barrier.

The impact of terrain-induced gravity waves (mountain waves) on orographic precipitation has received increased attention lately, since the vertical motions associated with these waves aloft can have a profound impact on the precipitation distribution. For example, Bruintjes et al. (1994) used radar observations to show how gravity waves over narrow ridges enhance precipitation upstream of the Mongollon Rim of Arizona. Meanwhile, Colle (2004) used two-dimensional simulations to show how the upstream tilt of a gravity wave over the crest can shift the precipitation distribution upstream for a relatively wide (50-km half-width) barrier under stably stratified conditions. More recently, Garvert et al. (2005, 2007) used aircraft in situ and radar observations as well as high-resolution (1.33-km grid spacing) simulations by the fifth-generation Pennsylvania State University–National Center for Atmospheric Research (PSU–NCAR) Mesoscale Model (MM5) to show the enhancement in vertical motions and precipitation by gravity waves over the narrow ridges of the windward Oregon Cascades during the 13–14 December 2001 event of the second Improvement of Microphysical Parameterization through Observational Verification Experiment (IMPROVE-2). They showed maxima in cloud water and snow production aloft to 6 km ASL associated with gravity waves produced by southwesterly flow interacting with the windward ridges of the Oregon Cascades. These vertical motion enhancements over the Cascades and European Alps are typically associated with enhancements in riming over a ridge (Medina and Houze 2003).

Smith and Barstad (2004) developed a linear theory of stable orographic precipitation, which included analytical solutions to the mountain-wave problem for different terrain geometries and appropriate cloud process time scales. This linear model has been successfully applied for barriers such as the Olympics (Smith and Barstad 2004), Oregon Cascades (Smith et al. 2005), and European Alps (Barstad and Smith 2005). The model produces precipitation variability over the windward slope as a result of gravity waves generated by the narrow ridges. For horizontal scales of 10–30 km, the maximum precipitation occurs near the ridge crests, since hydrometeor advection is compensated by any upstream wave tilt to the vertical motion.

It is clear that narrow windward ridges can enhance the precipitation locally over the ridge (Medina and Houze 2003; Garvert et al. 2007), but can a series of windward ridges increase the net precipitation along the windward slope? The answer is not obvious, since downwind of each ridge the precipitation may be reduced by downslope drying and removal of water vapor from the previous ridges. Garvert et al. (2007) addressed this question for their IMPROVE-2 case by replacing their real Cascade terrain in the MM5 with a smooth slope, while preserving the mean height of the Cascade crest. For a south-southwesterly flow at 1–2 km ASL, an air parcel over the Cascades likely encounters five to eight ridges that have 200–800-m-height perturbations before reaching the crest (see Fig. 1a of Garvert et al. 2007). They showed that for the 13–14 December IMPROVE-2 event the total windward precipitation at the surface was enhanced by 10%–15%.

More case studies are needed to further quantify the results of Garvert et al. (2007) for other synoptic settings and barrier dimensions. Meanwhile, additional motivation for these case studies can be obtained by completing a range of idealized studies, which is the goal of this study. The following motivational questions are addressed to improve our limited understanding of the impact of windward ridges on precipitation:

  • What is the impact of multiple ridges on the precipitation structures over a relatively wide barrier such as the Cascades?
  • How do the windward ridges affect the ability of the mountain to remove moisture and precipitate it, and what are the microphysical time scales?
  • How do the ridge impacts change for different flow speeds, stabilities, and freezing levels?

The next section discusses the model setup and parameters used to quantify the differences. Section 3 presents the results for three different cross-barrier flow speeds for weak, stably stratified conditions as well the impact of increased stratification, windward ridge height, and freezing level. The summary and conclusions are presented in section 4.

2. Data and methods

This study follows a similar approach outlined in Colle (2004), in which the MM5 was used in a two-dimensional configuration and initialized with a single sounding. For this study a 1500-gridpoint domain was applied with 1-km horizontal grid spacing, 38 vertical half sigma levels,1 and constant (fixed) lateral boundary conditions. The model top was set to 50 mb. The relatively large size of this domain combined with a short simulation (12 h) did not allow spurious reflections from the boundaries to impact the results. This was verified by comparing with some additional simulations using 3000 grid points (not shown), which yielded nearly identical results as the 1500-km grid. Additional runs were also completed using twice as many vertical levels, which also did not impact the results presented in this paper. Although a two-dimensional model has its limitations, it was more feasible than using a large three-dimensional domain, especially considering the 100–200 simulations that were completed. The 2D MM5 has been useful in studying downslope windstorms (Doyle et al. 2000), orographic precipitation (Colle 2004; Colle and Zeng 2004a, b), and microphysics (Rasmussen et al. 2002).

The MM5 simulations used a bell-shaped mountain ridge height (hm) of 2000 m and a half-width (a) of 50 km located at x = 525 km of the 1500-km domain. The terrain height, h(x),
i1520-0469-65-2-509-e1
and
i1520-0469-65-2-509-e2
includes a sinusoidal height perturbation (h′) for n number of windward ridges from the crest (x = 0) to −L upstream (Fig. 1), which was −3a (L = 150 km) for this study. Using this approach the average terrain height of the barrier does not change. The 50-km half-width and 2000-m mountain height parameters are similar to relatively wide barriers, such as the Oregon Cascades and California Sierras. The height perturbation (h′) was varied from 200, 400, and 800 m for n = 0, 1, 2, 4, 6, 8, 10, 12, 14, and 16 (Table 1). For n = 16 over a distance L = 150 km, there are ∼9 grid points across each windward ridge at 1-km grid spacing. A few additional simulations at 0.5-km grid spacing yielded very similar flow and precipitation results (within 5%); thus 1-km grid spacing was utilized in this study.

A uniform moist static stability (Nm), as defined in Durran and Klemp (1982) and Eq. (2) of Colle (2004), was initialized across the domain. A moist Nm was used, since the model was initialized as nearly saturated (98% relative humidity). An Nm of 0.005 and 0.01 s−1 was applied, with the weaker stability runs being similar to the low-level conditions observed during 13–14 December 2001 of IMPROVE-2 (cf. Fig. 5 of Garvert et al. 2005). An increase in stratification to Nm = 0.01 s−1 and variations in cross-barrier ambient flow (U) from 8, 15, and 30 m s−1 helped control the amount of flow blocking and gravity wave characteristics. Freezing levels (FLs) were specified to be either at 1000, 750, and 500 mb. As noted in Colle (2004), the precipitation amounts for the various Nm and FL settings cannot be compared quantitatively, since they have different integrated water amounts; however, the structure of the precipitation distribution and microphysical efficiencies can still be analyzed.

The MM5 was integrated for 12 h using the Thompson et al. (2004) microphysical scheme. This scheme includes six water species: water vapor, cloud water, rain, cloud ice, snow, and graupel. The Medium-Range Forecast (MRF) planetary boundary layer scheme was applied using a surface roughness length zo of 10 cm; however, no surface heat/moisture fluxes and radiation were included. Klemp and Durran’s (1983) upper-radiative boundary condition was used at the upper boundary to prevent gravity wave reflection. As described in Colle (2004), the 2D MM5 included the Coriolis effect for a representative midlatitude value of f ∼10−4 s−1.

A few quantities were calculated to better understand the precipitation generation and microphysical time scales. Precipitation efficiency has been used in previous studies to measure the ratio of precipitation fallout over the barrier to the amount of condensation and deposition aloft (Colle and Zeng 2004b; Colle 2004). However, this parameter is less useful for multiple windward ridges, since an air parcel condenses and evaporates several times while crossing multiple ridges, which can lead to artificially inflated condensation rate (Smith et al. 2003). A more useful parameter to measure the efficiency of the ridges to remove water vapor is the drying ratio (DR),
i1520-0469-65-2-509-e3
which measures the ratio of surface precipitation over a specified region (P) to the incoming water vapor (WV) flux (Smith et al. 2003). This was calculated for the MM5 by comparing the total surface precipitation accumulated for a region −3a upstream and +1a downstream of the crest (AB on Fig. 1) to the water vapor flux entering from the surface to the model top at −3a.2
Another way to quantify microphysical efficiency is to determine the characteristic residence time for all water and ice generated aloft to fall as precipitation (Smith et al. 2003). The residence time (RT) is defined within the box region as
i1520-0469-65-2-509-e4

3. Results

a. Ridge impacts for Nm = 0.005 s−1

The first series of ridge experiments (U8, U15, and U30) used an Nm of 0.005 s−1; U of 8, 15, and 30 m s−1; freezing level of 750 mb; and a ridge perturbation height of 400 m (Table 1). Figures 1 and 2 show the terrain and 6–12-h precipitation profiles for the U15 runs using n = 0, 1, 4, 8, 12, 16 windward ridges. For the n = 0 run (Fig. 1b), the precipitation maximum (26 mm) is located about ∼25 km upstream of the crest. When a single upstream ridge is used (n = 1), the precipitation distribution over the barrier is bimodal, with maxima of 20 and 27 mm located about 10 km upstream of the windward ridge and the crest, respectively. Meanwhile, a precipitation minimum (2–3 mm) exists within the 40-km-wide windward valley. For n = 4, the maximum precipitation is nearly 2–3 times larger than n = 0 over the windward slope. For each subsequent ridge along the windward slope in n = 4, the precipitation amount increases steadily from 22 mm ∼5 km upstream of first peak to 43 mm at ∼5 km downwind of the final windward peak. In contrast, the precipitation is less than 5 mm within the windward valleys. For n = 8 (Fig. 2b), the precipitation maxima over the ridges are 10%–20% less than n = 4. There is a ∼20% increase in precipitation from the first to the second ridge for n = 8, which suggests that the first ridge helps to enhance precipitation over the second ridge. Meanwhile, there is a slight decrease in the precipitation maxima over the next two subsequent ridges. This suggests that the precipitation over these first two ridges depletes some available moisture, which decreases the precipitation maxima immediately downstream, except over the upper windward slope, where the vertical motions are larger. For n = 12 and n = 16, the precipitation amount maximizes over the second ridge and just upwind of the crest. Interestingly, the precipitation over the windward ridges for n = 12 is greater than n = 16, suggesting that the n = 16 ridges are too narrow to enhance precipitation as efficiently as n = 12.

Figures 3 and 4 show cross sections of winds and precipitation hydrometeors for many of the above simulations. For n = 0 (Fig. 3a), there is a broad orographic snow cloud aloft associated with a single mountain wave aloft tilting upstream of the crest, while graupel exists over the upper windward slope. With the addition of a single ridge upstream of the crest (n = 1; Fig. 3b), the vertical motion with the upstream ridge results in a snow plume aloft (0.22 g kg−1) of similar magnitude to the snow above the crest, even though the crest is 1500 m higher. Graupel and rain rates are greater near the crest, since the effective slope of the ridge plus the bell-shaped barrier is steeper near the crest than over the lower windward slope. For n = 4 and 8 (Figs. 3c, 4a), the graupel amounts increase over each subsequent ridge toward the crest, while the percentage of snow increase toward the crest is less. For n = 12 and 16 (Figs. 4b,c), the perturbations in snow amount over the individual ridges is small aloft, as more snow is advected from one ridge to another rather than sublimating in between. However, the snow mixing ratios are still 30%–40% greater than n = 0. Meanwhile, graupel plumes below 3 km are still prevalent over the individual ridges for n = 12 and 16. For both n = 12 and 16, there is strong subsidence over the crest of the barrier, which advects hydrometeors to the surface, thus creating a localized surface precipitation maximum near the crest (Fig. 2b).

Figure 5a shows the average 6–12-h surface precipitation between points AB in Fig. 1b (50 km downwind of the crest to 150 km upstream of the crest) for the U8, U15, and U30 runs. For U15, there is an initial decrease from 12 to 10 mm from n = 0 to 1 associated with leeside subsidence downwind of the first windward ridge (Fig. 3b). Subsequently, the average windward precipitation increases nearly linearly to 16 mm for n = 12 (33% greater than n = 0), followed by a slight decrease for n > 12. The DR for AB increases from 0.34 at n = 0 to 0.45 at n = 12 (Fig. 5b). Therefore, the ridges increase the efficiency of water vapor removal and surface precipitation.

As the number of ridges is increased to 16, a line connecting the tops of the windward terrain peaks yields a higher and broader mountain (Fig. 4c). Perhaps the average precipitation enhancement for greater ridge numbers is simply the flow interacting with this somewhat higher and wider profile of ridge crests? Therefore, a separate n = 0 experiment was completed for U15 in which the barrier height and half-width was set to 2400 m and 85 km, respectively, which approximately follows the top of the ridge crests (U15ALT in Table 1). This higher and broader mountain enhances average precipitation between AB by ∼10% (Fig. 5a), but it is still less than half the U15 enhancement. Therefore, the circulation and microphysics over the individual ridges are more important in enhancing net windward precipitation than a silhouette topography that includes the tops of the ridge crests.

It was mentioned above that the first windward ridge may enhance the precipitation over the subsequent (second) ridge given their close proximity for n > 8. To test this, another experiment was completed identical to n = 12, except that only every other ridge was applied (N12ALT experiment). Figure 6 shows the precipitation profile across the barrier for this experiment as compared to n = 12. The precipitation over the first ridge is similar for both runs; however, the precipitation over the second and third ridges in N12ALT is 10%–20% less than the same ridges in n = 12. The precipitation over each subsequent ridge in N12ALT increases nearly linearly from the lower windward slope to the crest, while the n = 12 run has a larger precipitation increase from ridge number 1 to 2. Overall, the average precipitation between AB over the windward slope for N12ALT is ∼10% less than n = 12 (not shown), thus illustrating that the close proximity of the ridges helps to increase the net precipitation for the greater ridge numbers.

For the U15 runs the percentage of water vapor loss (WVL) upstream of the crest was separated into condensation and ice/snow deposition using output from the model every 15 min averaged between 6 and 12 h as in Colle and Zeng (2004a). The percentage of WVL from ice and snow deposition decreases from 25% to 8% as the ridge number is increased from 0 to 12, while the condensation increases to from 75% to 92% (not shown). However, a large fraction of this condensational gain is compensated by the evaporation between ridges, which increases from 5% of WVL for n = 0%–57% for n = 12, while there is little increase in sublimation.

The trend in average windward precipitation and DRs for U8 is similar to U15 for n = 1 to 12 (Fig. 5), with a ∼20% increase in precipitation and a DR increase from 0.30 to 0.37. For U30 (Fig. 5), the average precipitation increase (15%–20%) from n = 1 to 12 is less than the U15 runs, while the DR increases from 0.37 to 0.43 between n = 1 and 16. As will be highlighted below, the relatively short time for the flow to cross a ridge during the U30 runs does not allow time for hydrometeor growth and fallout as efficiently as U15, with the U15 runs having a larger DR than U30 for n > 8 (Fig. 5b). Meanwhile, the DRs and precipitation increase slightly for U30 from n = 12 to 16, since as the ridge spacing decreases there is less time for evaporation and sublimation between the ridges with the strong flow.

Figure 7 shows the maximum precipitation along the windward slope and how far upstream of the crest this maximum occurs for the various ambient wind speeds. For U15, the maximum precipitation is bimodal, with a peak of ∼45 mm for n = 4 and n = 12. The first maximum is ∼25 km upward of the crest, while the second is ∼17 km upwind. Meanwhile, the maximum precipitation for U8 occurs for n = 4 at ∼30 km upwind of the crest. For n > 4, the maximum U8 precipitation decreases and it is located 50 and 25 km upwind of the crest for n = 12 and n = 16, respectively. In contrast, the maximum precipitation for U30 increases nearly linearly from ∼60 to ∼72 mm for n = 2 to 10, while the maximum location shifts from 60 to 5 km upwind of the crest. These results suggest that the maximum precipitation can be increased by 60%–70% when just a few ridges are added under weak to moderate flow, while the maximum enhancement for U30 reaches its maximum for larger ridge numbers (n = 8–10).

The different precipitation distributions for the various wind speeds suggest that the vertical motions and associated microphysical processes differ. The vertical motion over the ridges is dependent on whether the gravity waves are vertical propagating or evanescent, as well as the mountain-wave structure over the full barrier. Vertical propagating waves are favored when the horizontal advection time or period of flow over terrain, T = L/U, where L is the terrain wavelength, is greater than the buoyancy period, 2π/Nm (Durran 2003). A shorter horizontal advection time than the buoyancy period does not allow much vertical oscillation; thus the wave becomes evanescent instead of vertically propagating. For the U15 runs, the minimum terrain wavelength to support vertically propagating waves (Lc) is ∼19 km, which occurs around n = 8 (Lc = 150 km/n), while for U8 and U30 the evanescent transition occurs for n = 16 and 4, respectively.

To show how the vertical motion amplitude changes as the precipitation becomes more enhanced with increasing ridge number, Fig. 8 presents the mass-weighted average of the vertical motion over the barrier to 50 mb for n = 0, 4, 8, and 12 of the U15 and U30 experiments. This was calculated by first adding the vertical motion multiplied by the corresponding pressure every 25 mb upward to the model top, and this total at each x grid point was divided by the sum of pressures at all levels. For n = 0 (Fig. 8a), there is a gradual increase in upward motion to 0.3 and 0.8 m s−1 over the windward slope for U15 and U30, respectively; however, weak subsidence exists ∼20 km upwind of the crest in the U15 run associated with the gravity wave tilting upstream of the crest with height (Fig. 3a). For n = 4 (Fig. 8b), the vertical motions are 2–3 times larger than n = 0. The horizontal wavelength of the mountain wave for n = 4 is less than the ridge spacing, resulting in upward motions of ∼0.5 m s−1 both over the windward and lower lee slopes of the ridges. Meanwhile, U30 consists of a single upward motion maximum of 1–2 m s−1 over the upper windward slopes that reaches a maximum around n = 8 (Fig. 8c). Recall that the ridge precipitation enhancement peaked for U30 at n = 8, while U15 peaked at n = 12 (cf. Fig. 5a). For n = 12 (Fig. 8d), the U15 vertical motions are more distinct over each ridge than n = 8 and are located just over the windward slopes. The vertical motions for n = 16 are slightly less than n = 12 for the U15 experiments (not shown). Overall, these results suggest that the maximum ridge enhancement on the average precipitation over the windward slope occurs when a distinct upward and downward motion couplet develops over each ridge, with minimal lee wave interference from the ridge upstream. This is favored as the gravity waves transition to more evanescent (n > 8 for U15 and n > 4 for U30), since the horizontal wavelength of the gravity wave is approximately that of the underlying terrain. When the gravity waves become more evanescent (n > 12 for U15 and n > 8 for U30), the net precipitation increase over the ridges slows or even decreases (Fig. 5a).

The microphysical impacts are also important for the ridge precipitation distributions. Figure 9 shows the vertically integrated snow, cloud water, and graupel averaged between AB for each wind speed and ridge number. For U8 and U15, there is a cloud water increase for n = 1 to 8 (Fig. 9a). Although cloud water decreases for U8 for n > 12, there is a compensating rainwater increase (Fig. 9b), resulting is little change in surface precipitation for n > 12. Therefore, much of the precipitation increase for U8 is associated with cloud water collision and coalescence processes at low levels given the limited snow growth aloft (Fig. 9c). In contrast, there is little cloud water increase for U30 with increasing ridge number, while there is enhanced graupel production for n = 1–4, which is within the vertically propagating gravity wave regime. Meanwhile, for U15 there is enhanced riming (graupel) for n = 1 to 8 (Fig. 9d), which also occurs with robust vertically propagating gravity waves. For both U15 and U30, there is 30%–50% increase in snow amounts over the barrier as the ridge numbers increase from 2 to 16 (Fig. 9c). This suggests that the vertical velocities produced aloft within the evanescent wave regime are still important in enhancing the snow growth at moderate to high wind speeds.

The hydrometeor residence time (RT) was calculated for each of the wind speed and ridge number (Fig. 10). For n = 0, the RTs range from 2300, 2100, and 2050 s for U8, U15, and U30, respectively, while the RTs are more similar (2050–2200 s) for n = 1. As n is increased further to n = 4, the residence time decreases as graupel and cloud water processes become more important (Fig. 9), with residence times ∼2050, 2000, and 1850 s for U8, U15, and U30, respectively. The longer residence time for slower wind speeds (U8) is associated with the slower precipitation growth rates during weaker orographic forcing. The U30 residence times are similar to U15 for n = 7 and 8, since U30 has more spillover of precipitation to the lee that sublimates.

b. Impact of increased stratification (Nm = 0.01 s−1)

To explore the impact of increased stratification, the stability was increased to Nm = 0.01 s−1, while the other parameters were the same as the U15 runs. Figure 11a shows the average precipitation between points AB for U = 8 m s−1 (N01U8), 15 m s−1 (N01U15), and 30 m s−1 (N01U30). For n = 0, the precipitation difference between N01U30 and N01U8 is greater than between U30 and U8. Colle (2004) conducted similar n = 0 simulations at 4-km grid spacing, and illustrated that for strong flow a greater upstream shift and increase in precipitation occurs when the stability is increased from 0.005 to 0.01 s−1 (U30 to N01U30). In contrast, the average precipitation over the barrier decreases when the stability is increased for weak to moderate flows (N01U15 and N01U8), since the greater stratification favors more low-level flow blocking and a shallow vertical wavelength for a linear hydrostatic mountain wave. The moist Froude number, U/(hmNm), decreases from 1.5 for N01U30 to 0.4 for N01U8; therefore, the surface flow in N01U8 is more blocked (<2 m s−1) over the lower barrier slope (cf. Fig. 2 of Colle 2004). The vertical wavelength of the mountain wave (2πU/Nm) in N01U8 is ∼6500 m, which yields a shallow snow area aloft below 3.5 km ASL (cf. Fig. 11a of Colle 2004).

All N01 simulations have a slight decrease in precipitation from n = 0 to n = 1 (Fig. 11a), given the subsidence in the lee of the windward ridge for n = 1 (not shown). The average precipitation across AB increases 28% from n = 2 to 6 for N01U30, while the increase is more gradual (∼12%) for N01U15 and very little for N01U8.

For N01U8 and n = 8 (Fig. 12a), there is some shallow precipitation enhancement over the ridges, but it is offset by minima in the valleys, thus resulting in little average precipitation enhancement (Fig. 11a). The orographic cloud for N01U15 is also limited to below 4 km ASL (Fig. 12b). Vertically propagating gravity waves are favored for N01U15, since the horizontal advection time or period over each ridge, L/U = 1250 s, where the windward ridge wavelength, L = 18.75 km, is greater than 2π/Nm (∼630 s). There are distinct oscillations associated with these waves aloft; however, the subsidence associated with the full-barrier mountain-wave tilting upstream of the crest creates downward motion above 5 km, resulting in little cloud above this level. This illustrates that the gravity wave response from the full barrier is important in determining how deep the ridges will produce precipitation aloft. This full barrier response is determined by the vertical wavelength (2πU/Nm), which increases as the flow increases and/or the stability decreases.

For N01U30, there are much deeper gravity wave perturbations over the windward slope, which results in deep plumes of graupel and snow to 6 km ASL for n = 8 (Fig. 12c). These vertical plumes of graupel become less pronounced for n = 12 for N01U30 (not shown); therefore, the precipitation decreases ∼10% for n = 6 to 12, and there is little increase for n > 12.

The DRs for the N01 runs are largest for U = 30 m s−1 (Fig. 11b), in which the DR increases from 0.38 for n = 0 to 0.44 for n = 6, and then gradually decreases to 0.42 for n = 12. The DRs for N01U30 are similar in magnitude to U30 (Fig. 5b), since both runs produce a deep orographic cloud aloft. In contrast, the weaker flow in the N01U15 and N01U8 runs generated a shallower vertical mountain wavelength and orographic cloud; thus less water vapor was converted to hydrometeors aloft and the DRs are much less than U15 and U8. For the N01U15 run, the DR of 0.26 for n = 0 and 2 increases slowly to 0.33 for n = 14 and 16. For N01U8, the DR only increases from 0.27 to 0.29 from n = 0 to 10, and then decreases to 0.26 for n = 16. The low DR and weak flow for N01U8 results in residence times that are fairly steady around 2900 s for all ridge numbers (not shown).

Some of the above results can be summarized by comparing the fractional enhancement of multiple ridge precipitation relative to the n = 0 run versus the moist Froude (Fr) number, U(hmNm) (Fig. 13). It was hypothesized based on recent field studies, such as for the Mesoscale Alpine Programme (MAP; Rotunno and Ferretti 2001), that if the flow is blocked for a Fr < 1, there would be less flow up and over the various windward ridges, and thus the potential for less precipitation enhancement over the barrier by the ridges. For relatively low moist Froude numbers (Fr ∼ 0.4) and n = 8, 12, and 16, the multiple ridges result in no average precipitation enhancement given the low-level flow blocking and weak mountain waves over the windward ridges (Fig. 12a). For moderate moist Froude numbers (0.8–1.5), there are more well-defined mountain waves and vertical motions over the barrier (Fig. 3), thus resulting in 20%–30% average enhancement for n > 8. When the moist Froude number is increased further to 3.0, the ridge enhancement is decreased to <20%, since the stronger cross-barrier flow favors less time for precipitation growth over the ridge and more spillover into the lee.

c. Impact of windward ridge amplitude

To explore the impact of windward ridge height on precipitation, the ridge amplitude (h′) was varied from 200 to 800 m for Nm = 0.005 s−1, U = 15 m s−1, and FL = 750 mb (Z200 and Z800 experiments). For n = 1 and 2 (Fig. 14a), there is little difference in average precipitation over the windward slope for the various windward ridge heights. For Z800 and small ridge number (n = 1–3), the precipitation increases over the individual ridges relative to Z400 and Z200 (not shown); however, the subsidence drying also increases over the valleys, which results in little net gain of average precipitation averaged over the windward slope. As n is increased from 4 to 12, the average precipitation over the windward slope increases more dramatically for Z800, reaching a maximum difference at n = 12. The precipitation increase for Z800 between n = 0 and 12 is associated with a 21% increase in average integrated cloud water between AB from 0.60 to 0.74 g kg−1, 15% increase in snow from 0.40 to 0.46 g kg−1, and an 81% increase in graupel from 0.11 to 0.20 g kg−1 (not shown). The microphysical residence times for Z800 decrease from 2000 s at n = 0 to 1650 s for n = 12, but then increased to 1800 s for n = 16 (not shown).

Overall, higher ridge heights have a more dramatic impact on average precipitation over the windward slope when the ridge number increases. However, even a series of relatively small (Z200) ridges can produce a 10%–15% enhancement in average precipitation over the windward slope (Fig. 14a). Most of the Z200 precipitation enhancement is associated with a cloud water increase at low levels from 0.59 to 0.65 g kg−1 (not shown), while there is little change in integrated snow and graupel given the shallow nature of the terrain perturbations (not shown).

The DR variations are similar to the change in the average precipitation over the windward slope (Fig. 14b). For the various ridge perturbation heights and n = 0–2, the DRs are 0.32–0.34. The DR for Z800 increases dramatically from 0.34 for n = 2 to 0.53 for n = 12, while the increase is limited to 0.46 and 0.37 for Z400 and Z200, respectively.

d. Impact of freezing level

Additional simulations were completed in which the FL was raised to 500 mb (FL500) and lowered to 1000 mb (FL1000) using U = 15 m s−1 and Nm = 0.005 s−1. For all freezing levels (Fig. 15a), the precipitation enhancement averaged over the barrier reaches a maximum of 15%–30% for n = 12. With the higher freezing level, nearly all the precipitation enhancement for FL500 is associated with a 10%–15% increase in cloud and rainwater over the barrier (not shown). The DR for the FL500 runs increases from ∼0.55 for n = 0–2 to 0.70 for n = 10 (Fig. 15b), so the ridges become very efficient in removing water vapor as the freezing level is raised. As a result of the greater warm-rain processes for FL500, the microphysical residence times for FL500 (1660 s for n = 0 and 1480 s for n = 16) are 15%–20% smaller than FL750 (U15).

For the 1000-mb FL (FL1000), the average precipitation enhancement between AB is ∼15% by n = 12 (Fig. 15a). There is a 30% increase in integrated cloud water from 0.20 to 0.30 g kg−1 from n = 0 to 12, but little increase in snow (not shown). The relatively low freezing level does not result in any rainwater increases, but the integrated graupel is increased from 0.04 to 0.12 g kg−1 (not shown), thus suggesting some snow riming by cloud water. When n is increased to 16, the snow is increased by 10%–15%, and there is a ∼10% decrease in graupel and cloud water. The reduction in riming for FL1000 yields microphysical residence times of 2300–2500 s, which are 30%–40% larger than FL500 (not shown).

Figure 16 shows the precipitation distribution for the FL500 and FL1000 experiments for n = 0, 4, 8, and 16. For FL500 (Fig. 16a), there are large precipitation maxima for n = 4, which are 3–4 times larger than n = 0 over the windward slope, with a maximum over the upper windward slope and crest. As the number of ridges is increased to 16, the precipitation over the lower windward slope becomes greater than the middle windward slope, since a large fraction of the available moisture is depleted downwind of the first series of ridges. In contrast, the precipitation maximum for FL1000 is shifted farther downwind of the first series of ridges over the lower windward slope as n is increased to 16. The ridges over the lower slope have less of an impact on the precipitation enhancement, since there is less riming than for FL750, while the snow is advected downwind of the first series of ridges. The DR for the FL1000 only increases slightly as the ridge number is increased, and it remains less than 0.20 (Fig. 15b).

4. Summary and conclusions

This paper explores the impact of multiple ridges along a broad windward slope on the flow and precipitation structures. This was accomplished by completing idealized two-dimensional simulations for a 2000-m mountain and 50-km half-width using MM5 for different stabilities, flow speeds, ridge numbers (n), and freezing levels. The results agree with a modeling case study from IMPROVE-2 (Garvert et al. 2007), in which relatively small ridges (200–400 m) were shown to have a profound impact on the precipitation distribution over the barrier. The impact of ridges on precipitation strongly depends on the number of ridges, with a few ridges (n < 4) favoring larger precipitation perturbations over the individual ridges, and greater ridge numbers (n > 8) favoring a 15%–35% increase the average precipitation over the windward slope. For light to moderate wind speeds (U = 8 and 15 m s−1), the multiple ridge impact on average precipitation peaks for n = 12 (ridge-valley wavelength of 12.5 km). For stronger cross-barrier flow (U = 30 m s−1), most of the average precipitation increase occurs around n = 8.

A separate n = 0 experiment for U15 was completed, in which the mountain height was raised from 2000 to 2400 m and half-width 50 to 85 km in order to mimic the flow over the terrain ridge tops of n = 16. The average windward precipitation only increases by ∼10%, which illustrates that the individual ridges are more important in enhancing the precipitation efficiency of the barrier than the higher silhouette topography the ridges create.

The precipitation differences for the various flows are related to the structure of the gravity waves over the windward peaks. A maximum precipitation response is favored when the horizontal wavelength produced by moist flow over the ridge is nearly equal to the ridge spacing, which is favored as the gravity waves transition from a vertically propagating to an evanescent regime (vertical frequency of flow over terrain nearly matches the static stability). A series of ridges results in multiple vertical motion enhancements, which increases the precipitation both locally and on average as the ridge wavelength becomes narrow enough, such that there is more limited sublimation and evaporation within the valleys. In addition, small ridge spacing is shown to have a synergistic effect on precipitation, in which a ridge over the lower windward slope helps increase the precipitation over the next subsequent ridge by 10%–20%.

For moist Froude numbers ≪1, the windward ridges do not increase the average precipitation over the windward slope. This is attributed to shallow flow blocking as well as the shallow smaller vertical gravity wavelength and associated orographic cloud over the crest. This result agrees with results from the MAP field project over the European Alps, in which there was little localized enhancement observed over the windward ridges when the low-level flow was blocked (Medina and Houze 2003). Meanwhile, these idealized results suggest that there may also be limited precipitation enhancement averaged over the windward ridges for large moist Froude numbers (Fr ≫ 1), since the stronger flow favors less time for precipitation growth over the individual ridges.

A higher freezing level (FL of 500 mb) creates large precipitation perturbations over the individual ridges because of enhanced riming and warm-rain processes. The percentage enhancement of precipitation over the windward slope is similar for a FL of 500 and 750 mb. However, there is a larger precipitation enhancement over the lower windward slope for multiple ridges for a higher freezing level given the warm-rain processes. In contrast, a lower FL (1000 mb) produces more precipitation enhancement for the n = 12–16 ridges shifted to the middle- and upper-windward slopes given the advection of ice aloft. These results also agree with MAP, where the largest precipitation maximum was located over the first few windward peaks when riming was occurring (Medina and Houze 2003).

The windward precipitation is also sensitive to the height of the ridge perturbations (h′). For small n (n < 4), the average precipitation over the windward slope does not increase with increasing h′ from 200 to 800 m. As n is increased from 4 to 12, the average precipitation over the windward slope increases more dramatically for h′ = 800 m, reaching a maximum difference at n = 12.

Future work will need to extend these idealized results to three dimensions as well as explore the impact of gravity waves on the microphysical characteristics across a series of narrow ridges using field datasets such as IMPROVE-2. In addition, the importance of mountain waves in enhancing orographic precipitation needs to be compared and combined with other recent hypotheses, such as increased microphysical growth produced by shear-induced turbulence over the windward slope (Houze and Medina 2005).

Acknowledgments

This research was supported by the National Science Foundation (ATM-0450444). The constructive comments from the three anonymous reviewers, which helped improve the manuscript, are appreciated. I thank Dr. Matthew Garvert for providing helpful suggestions during this study. Use of the MM5 was made possible by the Microscale and Mesoscale Meteorological (MMM) Division of NCAR, which is sponsored by the National Science Foundation.

REFERENCES

  • Alpert, P., , and H. Shafir, 1989: Mesoγ-scale distribution of orographic precipitation: Numerical study and comparison with precipitation derived from radar measurements. J. Appl. Meteor., 28 , 11051117.

    • Search Google Scholar
    • Export Citation
  • Barstad, I., , and R. B. Smith, 2005: Evaluation of an orographic precipitation model. J. Hydrometeor., 6 , 8599.

  • Bruintjes, R. T., , T. Clark, , and W. D. Hall, 1994: Interactions between topographic airflow and cloud/precipitation development during the passage of a winter storm in Arizona. J. Atmos. Sci., 51 , 4867.

    • Search Google Scholar
    • Export Citation
  • Colle, B. A., 2004: Sensitivity of orographic precipitation to changing ambient conditions and terrain geometries: An idealized modeling perspective. J. Atmos. Sci., 61 , 588606.

    • Search Google Scholar
    • Export Citation
  • Colle, B. A., , and Y. Zeng, 2004a: Bulk microphysical sensitivities within the MM5 for orographic precipitation. Part I: The Sierra 1986 event. Mon. Wea. Rev., 132 , 27802801.

    • Search Google Scholar
    • Export Citation
  • Colle, B. A., , and Y. Zeng, 2004b: Bulk microphysical sensitivities within the MM5 for orographic precipitation. Part II: Impact of barrier width and freezing level. Mon. Wea. Rev., 132 , 28022815.

    • Search Google Scholar
    • Export Citation
  • Doyle, J. D., and Coauthors, 2000: An intercomparison of model-predicted wave breaking for the 11 January 1972 Boulder windstorm. Mon. Wea. Rev., 128 , 901914.

    • Search Google Scholar
    • Export Citation
  • Durran, D., 2003: Lee waves and mountain waves. Encyclopedia of Atmospheric Sciences, J. R. Holton, J. A. Curry, and J. A. Pyle, Eds., Academic Press, 1161–1170.

    • Search Google Scholar
    • Export Citation
  • Durran, D., , and J. Klemp, 1982: On the effects of moisture on the Brunt-Väisälä frequency. J. Atmos. Sci., 39 , 21522158.

  • Garvert, M. F., , B. A. Colle, , and C. F. Mass, 2005: The 13–14 December 2001 IMPROVE-2 event. Part I: Synoptic and mesoscale evolution and comparison with a mesoscale model simulation. J. Atmos. Sci., 62 , 34743492.

    • Search Google Scholar
    • Export Citation
  • Garvert, M. F., , B. Smull, , and C. Mass, 2007: Multiscale mountain waves influencing a major orographic precipitation event. J. Atmos. Sci., 64 , 711737.

    • Search Google Scholar
    • Export Citation
  • Houze, R., , and S. Medina, 2005: Turbulence as a mechanism for orographic precipitation enhancement. J. Atmos. Sci., 62 , 35993623.

  • Klemp, J. B., , and D. R. Durran, 1983: An upper boundary condition permitting internal gravity wave radiation in numerical mesoscale models. Mon. Wea. Rev., 111 , 430444.

    • Search Google Scholar
    • Export Citation
  • Medina, S., , and R. Houze, 2003: Air motions and precipitation growth in alpine storms. Quart. J. Roy. Meteor. Soc., 129 , 345371.

  • Rasmussen, R. M., , I. Geresdi, , G. Thompson, , K. Manning, , and E. Karplus, 2002: Freezing drizzle formation in stably stratified layer clouds: The role of radiative cooling of cloud droplets, cloud condensation nuclei, and ice initiation. J. Atmos. Sci., 59 , 837860.

    • Search Google Scholar
    • Export Citation
  • Rotunno, R., , and R. Ferretti, 2001: Mechanisms of intense Alpine rainfall. J. Atmos. Sci., 58 , 17321749.

  • Sinclair, M. R., 1994: A diagnostic model for estimating orographic precipitation. J. Appl. Meteor., 33 , 11631175.

  • Smith, R. B., , and I. Barstad, 2004: A linear theory of orographic precipitation. J. Atmos. Sci., 61 , 13771391.

  • Smith, R. B., , Q. Jiang, , M. Fearon, , P. Tabary, , M. Dorninger, , J. Doyle, , and R. Benoit, 2003: Orographic precipitation and air mass transformation: An alpine example. Quart. J. Roy. Meteor. Soc., 129 , 433454.

    • Search Google Scholar
    • Export Citation
  • Smith, R. B., , I. Barstad, , and L. Bonneau, 2005: Orographic precipitation and Oregon’s climate transition. J. Atmos. Sci., 62 , 177191.

    • Search Google Scholar
    • Export Citation
  • Thompson, G., , R. M. Rasmussen, , and K. Manning, 2004: Explicit forecasts of winter precipitation using an improved bulk microphysics scheme. Part I: Description and sensitivity analysis. Mon. Wea. Rev., 132 , 519542.

    • Search Google Scholar
    • Export Citation
Fig. 1.
Fig. 1.

(a) Terrain profile (m) used in the idealized simulations for the 0, 1, and 4 windward ridge experiments for U = 15 m s−1. (b) The 6–12-h precipitation across the barrier for the 0, 1, and 4 windward ridge experiments for U = 15 m s−1. The dashed region AB in (a) represents the region where the average precipitation and drying ratio were calculated.

Citation: Journal of the Atmospheric Sciences 65, 2; 10.1175/2007JAS2305.1

Fig. 2.
Fig. 2.

(a), (b) Same as in Fig. 1 but for the 0, 8, 12, and 16 windward ridge experiments.

Citation: Journal of the Atmospheric Sciences 65, 2; 10.1175/2007JAS2305.1

Fig. 3.
Fig. 3.

Cross section for (a) 0, (b) 1, and (c) 4 windward ridges showing snow (gray every 0.04 g kg−1), graupel (dashed every 0.04 g kg−1), and rain (thin solid every 0.04 g kg−1), as well as potential temperature (solid every 4 K) and circulation vectors using the inset scale in (a).

Citation: Journal of the Atmospheric Sciences 65, 2; 10.1175/2007JAS2305.1

Fig. 4.
Fig. 4.

Same as in Fig. 3 but for (a) 8, (b) 12, and (c) 16 windward ridges.

Citation: Journal of the Atmospheric Sciences 65, 2; 10.1175/2007JAS2305.1

Fig. 5.
Fig. 5.

(a) Precipitation (mm) for the 6–12-h period vs ridge number averaged between points A and B in Fig. 1a for the U = 8 m s−1 (U8, dashed), 15 m s−1 (U15, black solid), and 30 m s−1 (U30, gray) experiments. The gray line is the average precipitation for the U15 run using a 2400-m mountain of 85-km half-width and n = 0. The wavelength of the windward ridges is also plotted along the x axis as well as the expected transition to evanescent gravity waves for the U30 (gray), U15 (solid black), and U8 (dashed). (b) Same as in (a) but for the drying ratio.

Citation: Journal of the Atmospheric Sciences 65, 2; 10.1175/2007JAS2305.1

Fig. 6.
Fig. 6.

The 6–12-h precipitation across the barrier for the n = 12 experiment for U = 15 m s−1 (gray lines) and an experiment (N12ALT in Table 1) in which the only every other ridge was used (black lines).

Citation: Journal of the Atmospheric Sciences 65, 2; 10.1175/2007JAS2305.1

Fig. 7.
Fig. 7.

Maximum surface precipitation (mm) accumulated during the 6–12-h period for any point between points A and B in Fig. 1a for the U = 8 m s−1 (U8, dashed), 15 m s−1 (U15, black solid), and 30 m s−1 (U30, gray) experiments. The plotted numbers show the distance (km) upstream from the crest where the maximum precipitation occurred.

Citation: Journal of the Atmospheric Sciences 65, 2; 10.1175/2007JAS2305.1

Fig. 8.
Fig. 8.

Mass-weighted vertical motion (solid, m s−1) averaged in the vertical vs distance for ridge numbers (a) 0, (b)4, (c) 8, and (d) 12 for the U15 (black) and U30 (gray) experiments in Table 1. Terrain is dashed for reference.

Citation: Journal of the Atmospheric Sciences 65, 2; 10.1175/2007JAS2305.1

Fig. 9.
Fig. 9.

Vertically integrated (a) cloud water, (b) rain, (c) snow, and (d) graupel (kg m−2) averaged between points A and B in Fig. 1a. The wavelength of the windward ridges is also plotted along the x axis in (a) as well as the expected transition to evanescent gravity waves for the U30 (gray), U15 (solid black), and U8 (dashed).

Citation: Journal of the Atmospheric Sciences 65, 2; 10.1175/2007JAS2305.1

Fig. 10.
Fig. 10.

Residence time (s) vs ridge number for all hydrometeors between points A and B for the U8 (dashed), U15 (solid black), and U30 (gray) experiments.

Citation: Journal of the Atmospheric Sciences 65, 2; 10.1175/2007JAS2305.1

Fig. 11.
Fig. 11.

(a) Precipitation (mm) for the 6–12-h period vs ridge number averaged between points A and B in Fig. 1a for Nm = 0.01 s−1 and U = 8 m s−1 (N01U8, dashed), 15 m s−1 (N01U15, black solid), and 30 m s−1 (N01U30, gray). (b) Same as in (a) but for the drying ratio.

Citation: Journal of the Atmospheric Sciences 65, 2; 10.1175/2007JAS2305.1

Fig. 12.
Fig. 12.

Cross section for n = 8 windward ridges for the (a) N01U8, (b) N01U15, and (c) N01U30 experiments showing snow (gray every 0.04 g kg−1), graupel (dashed every 0.04 g kg−1), and rain (thin solid every 0.04 g kg−1), as well as potential temperature (solid every 8 K) and circulation vectors using the inset scale.

Citation: Journal of the Atmospheric Sciences 65, 2; 10.1175/2007JAS2305.1

Fig. 13.
Fig. 13.

Fractional precipitation change as compared to the 0 ridge run vs moist Froude number for the 8, 12, and 16 ridge (see numbers on plot) simulations using the N01U8–N01U30 (gray lines) and U8–30 (black line) experiments. The precipitation is averaged between points A and B in Fig. 1a and accumulated for the 6–12-h simulation period.

Citation: Journal of the Atmospheric Sciences 65, 2; 10.1175/2007JAS2305.1

Fig. 14.
Fig. 14.

(a) Precipitation (mm) for the 6–12-h period vs ridge number averaged between points A and B in Fig. 1a for the ridge height experiments of Z = 200 m (Z200, dashed), 400 m (Z400, solid black), and 800 m (Z800, gray) experiments. (b) Same as in (a) but for the drying ratio.

Citation: Journal of the Atmospheric Sciences 65, 2; 10.1175/2007JAS2305.1

Fig. 15.
Fig. 15.

(a) Precipitation (mm) for the 6–12-h period vs ridge number averaged between points A and B in Fig. 1a for the freezing-level experiments of FL= 1000 mb (FL1000, dashed), 750 mb (FL750, solid black), and 500 mb (FL500, gray) experiments. (b) Same as in (a) but for the drying ratio.

Citation: Journal of the Atmospheric Sciences 65, 2; 10.1175/2007JAS2305.1

Fig. 16.
Fig. 16.

The 6–12-h precipitation (mm) across the barrier for the 0, 4, 8, and 16 windward ridge experiments for (a) FL500 and (b) FL1000.

Citation: Journal of the Atmospheric Sciences 65, 2; 10.1175/2007JAS2305.1

Table 1.

Parameters utilized for the various simulation experiments. For all experiments except U15ALT and N12ALT a set of runs was completed for a barrier with a 50-km half-width in which the windward ridge number (n) was increased from 0 to 16 (see text for details). The U15ALT used a 2400-m mountain with no windward ridges and an 85-km half-width, while the N12ALT is the same as n = 12 for U15 except that only every other ridge was used. The calculated vertically integrated water vapor flux was averaged for the 6–12-h period.

Table 1.
1

The 38 half sigma levels were σ = 0.997, 0.991, 0.985, 0.978, 0.971, 0.963, 0.954, 0.944, 0.933, 0.922, 0.910, 0.896, 0.881, 0.865, 0.848, 0.829, 0.808, 0.786, 0.763, 0.737, 0.710, 0.681, 0.650, 0.617, 0.583, 0.546, 0.507, 0.468, 0.427, 0.384, 0.341, 0.297, 0.253, 0.209, 0.167, 0.126, 0.086, and 0.049.

2

Smith et al. (2003) defined the box for DR to the cloud top of 5 km, while in this study it was defined to the model top, since the cloud depth varied substantially between runs. Since approximately 90% of the moisture flux is below 5 km, there is little difference between the two approaches.

Save