Precipitation Patterns in The Lee of The Santa Cruz Mountains of California

Chris Ineich

Writer’s comment: I first learned about orographic uplift and the rain shadow effect from watching a TV news weather segment. These con-cepts explained the dramatic differences in rain totals between the very wet Santa Cruz Mountains and the comparatively arid Santa Clara Valley. They also confirmed my own weather observations. I had begun to measure rainfall with a coffee can and ruler in my backyard in the Santa Cruz Mountains and noticed the rain totals I collected were consistently greater than those in the nearby flatlands around the bay. I was convinced my elevated rain totals resulted from a local orographic effect.
         Years later, I came across an article by another young weather enthusiast that forced me to rethink my assumptions. The author hypothesized it was possible for the orographic effect to increase precipitation for a large distance downwind of mountains and offered observational evidence to support this idea. I was intrigued—this theory defied everything I had read in textbooks about lee-side rain shadows. I set up my own rain collection study modeled after the article’s. I used the statistics I learned in a college course to analyze the data, and the results were more consistent than I expected. This project was my first attempt at basic research and broadened my perceptions of the impacts of mountainous terrain on precipitation patterns.
—Chris Ineich

Instructor’s comment: One of my favorite things about English 104E, Writing in the Sciences, is that students can write about their own research. I was intrigued by Chris’s research on rainfall patterns. He had read information about rainshadows that contradicted what he had learned about rainfall patterns. But instead of dismissing the author’s theories, he decided to test them. To Chris’s surprise, his results confirmed the research. When I assigned a report on field research, Chris did some reading on rainfall to put his work into context then reanalyzed his data. The result is the analytical report you see here; weather buffs and outdoors people will find it fascinating.
—Margaret Eldred, English Department

The factors influencing precipitation patterns in the lee of the Santa Cruz Mountains of California were studied. The variables analyzed included wind speed, wind direction, storm intensity, topography, and distance downwind from the mountains. Five standard rain gauges were set up along a SSW-NNE transect from the crest of the northern Santa Cruz Mountains to approximately eleven kilometers downwind and data were collected for five major storm events. A strong correlation was found between precipitation and downwind distance for four of the five storm events. A somewhat weaker precipitation-distance correlation that resulted for one precipitation event was likely a result of the showery nature of the storm and light winds. An index of storm intensity was developed for the study and a moderately strong correlation was found between lee-side precipitation totals and storm intensity. High intensity storms tended to produce smaller declines in precipitation from mountain to lee region than low intensity rain events. No relationship was found between lee-side topography and precipitation patterns.

         The effects of mountainous terrain on precipitation patterns can be profound. When moist air is forced to rise over elevated terrain the resulting cooling reduces the capacity of the air to hold water vapor and condensation occurs. This process is called the orographic effect and can result in much elevated precipitation on the windward slopes of mountains. The enhanced condensation and precipitation on windward slopes reduces the water vapor load of the atmosphere leading to reduced precipitation in the “rain shadow” region of the mountain’s lee.
         Scientists have long been fascinated with the impacts of elevated topography on precipitation patterns. Orographic precipitation phenomena can certainly be dramatic. The world’s rainiest climates are found on the windward slopes of major mountain ranges while the most arid deserts often lie in the lee of the same mountain complexes. With such a dramatic effect on regional climates and biomes, it is no surprise that so many researchers have chosen to study this topic. Atmospheric scientists, hydrologists, ecologists, geographers and soil scientists have taken great interest in orographic precipitation processes.
         Recent contributions to the study of orographic precipitation have come from around the world and highlight some of the critical interactions that occur between moist air masses and complex terrain. Working in Switzerland, Gysi (1996) found that the distribution of orographically influenced rainfall was strongly dependent on the direction from which a storm moved into a region of elevated topography. Those slopes that faced in the direction of an approaching storm received more rain than did terrain located downwind.
         Another study in England revealed the influence of wind speed on orographic enhancement with the greatest precipitation events associated with the strongest winds (Hill, Browning, and Bader, 1981). A similar study in Scotland verified the relationship between strong winds and an enhanced orographic effect (Weston and Ray, 1994). Rainfall increased almost linearly with increasing wind speed over the first area of high ground encountered in a storm’s path.
         Despite decades of intensive research, orographic precipitation processes remain among the least understood of atmospheric phenomena. Much of the research has focused on only a few of the world’s most prominent mountain belts, leaving the vast majority of mountainous regions unstudied. The potential effects of elevated topography on precipitation processes are as diverse as the terrain itself, and findings in one mountain region may not carry over well to another area of complex terrain. The data gathering process is also hampered by the inaccessibility of many remote mountainous regions.
         The most significant gap in our understanding of orographic precipitation phenomena concerns that precipitation which falls in the lee of mountain barriers. Limited data suggest that precipitation events in the lee of mountain ranges may be even more variable than on windward slopes, yet precipitation research in these “rain shadow” regions is particularly sparse.
         Only two recent studies have focused on lee-side precipitation. Chater and Sturman (1998) studied the conditions that favored the “spillover” of precipitation to the lee of the Southern Alps. The researchers noted the importance of wind speed, latent instability, and frontal intensity in determining the significance of lee-side precipitation events.
         With limited data, Haugland (1998) concluded that lee-side precipitation was highly dependent on wind speed and storm intensity. He found that storms with considerable precipitable water and strong winds produced major rain totals in lee-side areas because the moisture-laden clouds were rapidly carried downwind in their orographically lifted positions.Weak winds and little precipitable water produced storms that “rained out” on windward slopes, leading to rapid sinking and drying of the air mass on the lee side.
         To test the hypotheses of Haugland and to expand our general understandings of lee-side precipitation patterns, we conducted a field observational study of precipitation in the lee of the Santa Cruz Mountains of California. We examined the influence of several variables on precipitation events including wind speed and direction, atmospheric instability, cloud water content, lee-side topography, and distance from the crest of the mountain barrier. Precise measurements for some of these variables, such as instability and cloud water content, were not available and sound approximations were made when necessary. These are noted in the text.

Study Site
         The field data were collected in the lee of the northern Santa Cruz Mountains, a sub-range of the California Coast Ranges (Figure 1). The range spans approximately 80 km, from just south of San Francisco to west of Gilroy, and runs more or less in a north-northwest to south-southeast direction. To the west of the Santa Cruz Mountains lies the Pacific Ocean, and to the east is San Francisco Bay in the north and the Santa Clara Valley in the south. The crest of the range reaches a maximum elevation of 1137 meters at Loma Prieta near its southern terminus and averages about 550 meters in height. The elevation of the crest along the western margin of our study transect is approximately 690 meters.
         Despite the modest height of the Santa Cruz Mountains, orographic enhancement of precipitation over much of the range is substantial and a pronounced rain shadow exists in its lee. Long-term precipitation data is limited but suggests that average annual rainfall in the more orographically favored areas exceeds 1500 mm (Gilliam, 1962). No long-term rainfall data was available for the segment of the range considered in this study, but we conservatively estimated it to be between 1000 and 1300 mm/yr along the crest. Precipitation near San Francisco Bay, approximately 15 to 20 km downwind from the crest of the mountains, averages about 500 mm/yr.
Figure 1. Map showing the locations of the rain gauges used in the study: x1-El Corte de Madera, x2-Ca–ada Road, x3-Pulgas Ridge, x4-Crestview, x5-Belmont.

         An important consideration of the study was the role of lee-side topography on precipitation patterns in our study region. The lee of this part of the Santa Cruz Mountains is broken into hills and valleys that we thought might impact precipitation on a very local scale. Our study transect traversed the deep valley that marks the course of the San Andreas fault and climbed over Pulgas Ridge, one of the prominent ridgelines of the eastern San Francisco peninsula. Our Crestview study site sits atop the high point of this ridge at approximately 270 meters. East of Pulgas Ridge are several hills of less than 200 meters height. The northeastern most of our study points, Belmont, was located among these low hills. Figure 2 shows a topographic cross section of the study transect and the locations of each of the study sites relative to the major terrain features.
Figure 2. A topographical profile of the study transect.

Materials and Methods
         The design of our study was closely modeled after the work of Haugland (1998). The materials consisted of five identical standard rain gauges made by Taylor. The precision of the measurements taken between the gauges was verified by pouring measured amounts of water among the gauges and noting the equivalent readings.
         A study transect, located along a south-southwest to north-northeast orientation, was located in the lee of the northern Santa Cruz Mountains. The orientation was chosen because it parallels the average direction of the wind during most major storm events in this area. Because this orientation reflects only the average wind direction during storms, a certain amount of error was expected in our results depending on how significantly the winds during particular storm events deviated from the average.
         The five rain gauges were located along the pre-selected route beginning at the crest of the Santa Cruz Mountains and extending downwind for approximately 11 kilometers (Figure 1, Table 1). The exact locations of the rain gauges reflect a compromise between our desire to maintain the integrity of the study transect, the need to respect private property rights, and our interest in reducing the risk of vandalism that would have sacrificed the quality of our results. Four of the gauges were located within public wildlands in areas we felt would be well concealed from potential vandals. The fifth gauge was located in the backyard of the author. Every attempt was made to locate the gauges away from obstructions that might impact the rain catch during windy conditions; however, some error in this regard was unavoidable. All precipitation measurements were made by the same individual to minimize the error associated with manually read rain gauges.

Table 1. Approximate distance in kilometers from the crest of the Santa Cruz Mountains and height in meters of the study sites.

Rain gauge location Approximate distance from mountain crest (km) Approximate height above sea level (m)
El Corte de Madera 0.0 690
Cañada Road 6.0 120
Pulgas Ridge 7.5 165
Crestview 9.0 270
Belmont 11.0 45

         Precipitation totals were collected shortly after each precipitation event during the study period, between February 1999 and November 2000. Average wind speed and direction and maximum wind speed and direction were determined for each storm event by examining hourly data logged at a remote, automated weather station operated by the National Weather Service and located near our Pulgas Ridge study site (Pulgas RAWS).The precipitation and wind data are presented in Table 2.

Table 2. Precipitation totals (mm) and wind data (m/s) for each of five storm events.

  Storm 1 Storm 2 Storm 3 Storm 4 Storm 5
Mean wind speed and direction for storm (m/s) SSW 8 W 1 SW 4 WSW 2 SSW 5
Max. wind speed and direction for storm (m/s) SSW 17 NW 8 SSW 15 SW 13 SW 17
Precipitation (mm)
El Corte de Madera M* 37 49 51 33
Canada Road 55 15 39 35 19
Pulgas Ridge 52 15 39 27 15
Crestview 48 17 36 25 15
Belmont 43 19 34 23 11
*Missing Data

         We were interested in determining the role of atmospheric instability and cloud water content on precipitation in our study area, but precise data for these parameters were lacking. We rationalized that a sound approximation of these variables could be made by calculating the mean total precipitation received among all the rain gauges for each storm. Greater instability and cloud water content produce storms that dump more precipitation. With this in mind, we combined the instability and cloud water variables into a generalized parameter we termed storm intensity. We felt confident in making this generalization because of the qualitative rather than precise quantitative nature of the research.
         Correlation and least squares regression analysis were performed to highlight the significance of the relationships between precipitation events and the study variables. Table 3 depicts these values.

Table 3. Correlation and least squares regression analysis of precipitation to downwind distance data for five storm events.

  r r2 y = a + bx
Storm 1 0.999 0.997 69.8 - 2.4x
Storm 2 0.807 0.651 32.7 - 1.8x
Storm 3 0.989 0.979 48.6 - 1.4x
Storm 4 0.983 0.967 50.3 - 2.7x
Storm 5 0.986 0.973 32.1 - 2.0x

         The precipitation data for each storm are presented in Table 2. Our data clearly showed a linear relationship between storm totals and distance from the crest of the Santa Cruz Mountains.The calculated correlation values for each of storms 1, 3, 4, and 5 all exceeded .980 indicating a very strong linear relationship. The correlation for storm 2 was somewhat weaker than the others.
         A similar, although less strong, correlation was noted between storm intensity and lee-side precipitation (Table 4). High intensity storms tended to produce less significant declines in precipitation from mountain summit to lee region as compared to lower intensity storms. The storm of lowest intensity resulted in the largest percent decline in precipitation from mountain crest to lee. The most intense storm by our definition, however, produced the second smallest decline in precipitation . Of the five storms analyzed, the smallest percent decline in precipitation occurred with the storm we defined to be the second most intense. No pattern was observed in the data between precipitation totals and the topographic variations of the lee region under study.

Table 4. Relationship between storm intensity and precipitation reaching lee region as calculated for five storm events.

  Storm 1 Storm 2 Storm 3 Storm 4 Storm 5
Storm intensity (mean precipitation for all study sites) 54 21 39 32 19
Mean precipitation for lee-side sites 50 17 37 28 15
% precipitation from mountain crest reaching lee region 71 46 76 55 45
% decline in precipitation from mountain to lee 29 54 24 45 55
Correlation between storm intensity and precipitation reaching lee region 0.871        

         Our results provide considerable support to the hypotheses outlined by Haugland (1998). Haugland stressed the significance of wind speed, wind direction, cloud water content, and atmospheric instability on lee side precipitation patterns. Our findings suggest that these variables affect lee-side precipitation to the same degree as they affect windward slope precipitation.
         The strong linear relationship between lee-side precipitation totals and distance downwind from the crest of the Santa Cruz Mountains is made evident by the values in Figure 3. Most of the discrepancies in the correlation values away from the linear can be well explained based on wind speed and direction. The remarkable precipitation-distance correlation for storm 1 is likely a product of the strong and sustained south-southwesterly winds during that particular precipitation event. Haugland predicted that strong winds would carry orographically lifted clouds into the lee region rapidly enough to enhance precipitation for a distance downwind before the force of gravity pulled the clouds back to their un-lifted level. With a constant wind, we expect this situation to produce a linear decline in precipitation from mountain to lee region. Storm 1 clearly exemplifies this model.

Figure 3. Scatterplot of Storm 1 precipitation vs. distance downwind from the mountains with associated best fit line determined by linear regression analysis.

         The much weaker precipitation-distance correlation evident for storm 2 is also explainable. The winds during this storm were the lightest of the five precipitation events analyzed. Light winds minimize the downwind enhancement of orographic precipitation producing generally light precipitation amounts in the lee region. The erratic precipitation totals measured at the lee-side sites probably reflected the very showery nature of the system and not the influence of any of the variables we studied.
         Each of storms 3, 4, and 5 produced moderately strong precipitation-distance correlations reflecting their significant south-southwesterly winds. The correlations for these storms are somewhat less strong than for storm 1, probably as a result of minor deviations in wind directions from the SSW. A typical mid-latitude cyclone produces a succession of wind directions in passing a given location. Any precipitation that fell while the wind was blowing from a direction other than that which paralleled our study transect was expected to produce data inconsistencies. With enough wind data it would be possible to isolate that precipitation that falls with any given wind direction; the resulting precipitation data would likely produce very strong correlations.
         Our findings also provide some evidence that the percent decline in precipitation from the crest of the Santa Cruz Mountains to the lee region is dependent on storm intensity but further studies on this relationship are necessary. We expected that a stable atmosphere and storm clouds of low water content would combine to produce major precipitation declines from mountain to lee region. Because the stable atmosphere limits the amount of orographic lifting that can be realized on the windward slopes, the modest height of the storm clouds is quickly lost on the lee side. The very limited water supply of such storms is also likely to be quickly depleted on the windward slopes, even with the weak orographics.
         The calculated storm intensity—lee-side precipitation correlation was not as strong as we hoped to find, probably because of errors in our storm intensity measurements. We assumed that storm events accompanying a more unstable atmosphere and with high cloud water contents would always produce greater precipitation totals over our study area than storm events with less instability and cloud water content. This is not always the case, however. Other variables such as wind and temperature can greatly influence the distribution of precipitation over a given region. A very dense network of stations continually monitoring all of these parameters during the passing of a storm would be necessary for truly accurate calculations. The limited budget and manpower of this study made the establishment of such a dense network of monitoring stations an impossibility, however.
         We found no evidence that the topographic variations of our lee-side study area has any effect on precipitation patterns. We suspected that the 270 meter tall Pulgas Ridge that bisected our study transect might impact precipitation on a very local basis. No such local impact is evident from our data, and the scatterplots presented in Figure 3 show a steady decline in precipitation across the ridge for four of the five storm events.
         Precipitation actually increased across the ridge during storm 2 with a lee-side precipitation maximum recorded at Belmont. The Bel-mont site was the lowest in elevation of our study locations and the most distant from the Santa Cruz Mountains, leading us to suspect that this event was an anomaly resulting from the showery distribution of the precipitation and not a local orographic effect. It is likely that for storms with winds sufficient to produce an orographic effect, the potential orographic influence of the elevated lee-side topography is eliminated because of the much greater lift already provided by the neighboring Santa Cruz Mountains. The blocking effect of the Santa Cruz Mountains may also reduce low level wind speeds in the lee region to the point that orographic lifting can no longer be realized. This situation would lead to an obvious enhancement of precipitation along the crest of the Santa Cruz Mountains followed by a steady decline in precipitation across the lee region regardless of the shape of the topography.
         It seems possible, however, that a low pressure system deep enough to produce strong southeasterly winds might result in local orographic lifting along the eastern slopes of Pulgas Ridge. In this case we would expect to see a local precipitation maximum at the elevated Crestview site and a decline westward until additional lifting was provided by the Santa Cruz Mountains. Additional research is necessary to test this hypothesis.


Chater, AM; Sturman, AP. (1998) Atmospheric conditions influencing the spillover of rainfall to lee of the Southern Alps, New Zealand. International Journal of Climatology. Vol 18, No 1, pp 77-92.

Gilliam, H. (1962) Weather of the San Francisco Bay Region. Berkeley: University of California Press.

Gysi, H. (1998) Orographic influence on the distribution of accumulated rainfall with different wind directions. Atmospheric Research. Vol 47-48, pp 615-633.

Haugland, M. (1998) The rain shadow effect, accessed 2/10/00.

Hill, FF; Browning, KA; Bader, MJ. (1981) Radar and rain gauge observations of orographic rain over South Wales. Quarterly Journal of the Royal Meteorological Society. Vol 107, No 453, pp 640-670.

Weston, KJ; Roy, MG. (1994) The directional dependence of the enhancement of rainfall over complex orography. Meteorological Applications. Vol 1, No 3, pp 267-275.