• Anderson, B. T., , Roads J. O. , , and Chen S. C. , 2000: Large-scale forcing of summertime monsoon surges over the Gulf of California and the southwestern United States. J. Geophys. Res., 105 , 2445524467.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Balling R. C. Jr., , , and Brazel S. W. , 1987: Diurnal variations in Arizona monsoon precipitation frequencies. Mon. Wea. Rev., 115 , 342346.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Becker, E. J., , and Berbery E. H. , 2008: The diurnal cycle of precipitation over the North American monsoon region during the NAME 2004 field campaign. J. Climate, 21 , 771787.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Berbery, E. H., , and Fox-Rabinovitz M. S. , 2003: Multiscale diagnosis of the North American monsoon system using a variable-resolution GCM. J. Climate, 16 , 19291947.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Dixon, P. G., 2005: Using sounding data to detect gulf surges during the North American monsoon. Mon. Wea. Rev., 133 , 30473052.

  • Dixon, P. G., 2006: An ALARMS climatology of Gulf of California moisture surges and the associated precipitation, 1993–2004. Phys. Geogr., 27 , 86108.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Douglas, M. W., , and Leal J. C. , 2003: Summertime surges over the Gulf of California: Aspects of their climatology, mean structure, and evolution from radiosonde, NCEP reanalysis, and rainfall data. Wea. Forecasting, 18 , 5574.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Douglas, M. W., , Maddox R. A. , , Howard K. , , and Reyes S. , 1993: The Mexican monsoon. J. Climate, 6 , 16651677.

  • Freidlin, B., , and Gastwirth J. L. , 2000: Should the median test be retired from general use? Amer. Stat., 54 , 161164.

  • Fuller, R. D., , and Stensrud D. J. , 2000: The relationship between tropical easterly waves and surges over the Gulf of California during the North American monsoon. Mon. Wea. Rev., 128 , 29832989.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Hales, J. E., 1977: Relationship of convective cooling to nocturnal thunderstorms at Phoenix. Mon. Wea. Rev., 105 , 16091613.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Higgins, R. W., , and Shi W. , 2005: Relationships between Gulf of California moisture surges and tropical cyclones in the eastern Pacific basin. J. Climate, 18 , 46014620.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Higgins, R. W., , Shi W. , , and Hain C. , 2004: Relationships between Gulf of California moisture surges and precipitation in the southwestern United States. J. Climate, 17 , 29832997.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • King, T. S., , and Balling R. C. , 1994: Diurnal variations in Arizona monsoon lightning data. Mon. Wea. Rev., 122 , 16591664.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Kuwagata, T., , Numaguti A. , , and Endo N. , 2001: Diurnal variation of water vapor over the central Tibetan Plateau during summer. J. Meteor. Soc. Japan, 79 , 401418.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • McCollum, D. M., , Maddox R. A. , , and Howard K. W. , 1995: Case study of a severe mesoscale convective system in central Arizona. Wea. Forecasting, 10 , 643665.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Moore, D. S., 1977: Generalized inverses, Walds method, and construction of chi-squared tests of fit. J. Amer. Stat. Assoc., 72 , 131137.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Reiter, E. R., , and Tang M. C. , 1984: Plateau effects on diurnal circulation patterns. Mon. Wea. Rev., 112 , 638651.

  • Rodgers, J. L., , and Nicewander W. A. , 1988: Thirteen ways to look at the correlation coefficient. Amer. Stat., 42 , 5966.

  • Svoma, B. M., , and Balling R. C. Jr., 2009: An anthropogenic signal in Phoenix, Arizona winter precipitation. Theor. Appl. Climatol., 315321.

    • Search Google Scholar
    • Export Citation
  • Wallace, J. M., 1975: Diurnal variations in precipitation and thunderstorm frequency over the conterminous United States. Mon. Wea. Rev., 103 , 406419.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • View in gallery

    The 156 COOP stations examined. Shown here are the average annual precipitation for July and August (shaded dots) and the fraction of the total presurge and postsurge precipitation events that occurred during the postsurge periods (open circles). The rough location of the Sierra Madre Occidental is shown in the inset as SMO.

  • View in gallery

    Locations of the 63 hourly rain gauges analyzed around the Phoenix metropolitan area (dots). The percentage of nocturnal rain events (events occurring between 2000 and 0800 LST) during surge periods are represented by open-circle diameter. White (black) dots represent positive (negative) values, which indicate greater (lesser) percentages of nocturnal rainfall events during surge periods.

  • View in gallery

    Time of maximum for stations for which the first harmonic was calculated, portrayed using “wind direction” symbols with the usual meteorological definition. Flags with open (filled) circles indicate time of maxima for surge (nonsurge) periods. “Northerly” flags indicate 0000 LST, “easterly” flags indicate 0600 LST, etc.

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 6 6 1
PDF Downloads 5 5 0

The Influence of Monsoonal Gulf Surges on Precipitation and Diurnal Precipitation Patterns in Central Arizona

View More View Less
  • 1 School of Geographical Sciences and Urban Planning, Arizona State University, Tempe, Arizona
© Get Permissions
Full access

Abstract

North American gulf-surge events, northward low-level influxes of cool and moist air from the Gulf of California, were statistically related to monsoonal precipitation and the associated diurnal cycle for the state of Arizona. Using Dixon’s Assessing Low-Level Atmospheric Moisture using Soundings (ALARMS) method as an indicator for gulf surges, a sequence of surge events was identified for the months of July and August for the period from 1957 to 2008. A network of Arizona precipitation gauges was stratified by the surge events occurring over this period. The findings indicate that gulf surges accounted for a significant majority of rainfall events in Arizona. This signal was most apparent in the drier central and southwestern deserts (including the Phoenix metropolitan area) and least apparent in the wetter eastern and southeastern portions of the state. Diurnal patterns in rainfall were identified for the Phoenix metropolitan area and its surroundings. A strong diurnal cycle was apparent in precipitation associated with both surge and nonsurge periods over the Phoenix area, with a greater tendency toward nocturnal precipitation during gulf-surge events. These findings suggest that dissipating afternoon thunderstorms east and northeast of the Phoenix area act as catalysts for the nocturnal storm development that is prevalent in this area.

Corresponding author address: Bohumil Svoma, 2750 S. Morrow St., Tempe, AZ 85282. Email: bohumil.svoma@asu.edu

Abstract

North American gulf-surge events, northward low-level influxes of cool and moist air from the Gulf of California, were statistically related to monsoonal precipitation and the associated diurnal cycle for the state of Arizona. Using Dixon’s Assessing Low-Level Atmospheric Moisture using Soundings (ALARMS) method as an indicator for gulf surges, a sequence of surge events was identified for the months of July and August for the period from 1957 to 2008. A network of Arizona precipitation gauges was stratified by the surge events occurring over this period. The findings indicate that gulf surges accounted for a significant majority of rainfall events in Arizona. This signal was most apparent in the drier central and southwestern deserts (including the Phoenix metropolitan area) and least apparent in the wetter eastern and southeastern portions of the state. Diurnal patterns in rainfall were identified for the Phoenix metropolitan area and its surroundings. A strong diurnal cycle was apparent in precipitation associated with both surge and nonsurge periods over the Phoenix area, with a greater tendency toward nocturnal precipitation during gulf-surge events. These findings suggest that dissipating afternoon thunderstorms east and northeast of the Phoenix area act as catalysts for the nocturnal storm development that is prevalent in this area.

Corresponding author address: Bohumil Svoma, 2750 S. Morrow St., Tempe, AZ 85282. Email: bohumil.svoma@asu.edu

1. Introduction

The influx of moisture associated with the North American monsoon during the months of June–September accounts for the majority of annual precipitation in much of Arizona, contributing up to 60% in the southern portion of the state (Douglas et al. 1993). Although some researchers have found that much of this precipitation is linked to gulf-surge events (Dixon 2006; Becker and Berbery 2008), which are low-level influxes of cool and moist air from the Gulf of California northward into northern Mexico and Arizona, the findings of other investigators are more ambiguous in this regard (Douglas and Leal 2003; Higgins et al. 2004). The passage of easterly waves (Fuller and Stensrud 2000) and tropical cyclones approaching the Gulf of California (Anderson et al. 2000; Higgins and Shi 2005) are recognized as typical catalysts for gulf surges (Becker and Berbery 2008).

Monsoonal rainfall in Arizona is characterized by a prominent diurnal cycle with a maximum occurring in the late afternoon for much of the state (Balling and Brazel 1987). An exception to this is the nocturnal maximum for rainfall and thunderstorms over the Phoenix metropolitan area and its immediate surroundings (Wallace 1975; Hales 1977; Reiter and Tang 1984; Balling and Brazel 1987; King and Balling 1994). Possible mechanisms for the nocturnal maximum in storm activity for this region include 1) the destabilization of the midtroposphere during the evening hours as a result of westward-moving evaporatively cooled air from afternoon thunderstorms in the higher terrain to the east and northeast of the Phoenix area, 2) nocturnal surface convergence due to downslope winds, 3) the tendency for convective systems to propagate along cold outflows from higher to lower elevations (King and Balling 1994), and 4) thermally induced daytime mountain–valley circulations that result in enhanced subsidence and thus daytime rainfall suppression over the valley, which is common over land in complex terrain during the warm season (Kuwagata et al. 2001).

Studies using a long climatology of both gulf-surge events and hourly rainfall data are relatively limited in the literature regarding the influence of gulf surges on precipitation in Arizona. Higgins et al. (2004) studied the influence of these surges on precipitation events in the southwestern United States by using a climatology of surge events from 1977 to 2001 based on surface data from Yuma and Tucson, Arizona. Higgins et al. (2004) related the occurrence of gulf surges to precipitation totals at a resolution of 1° latitude by 1° longitude using the Climate Prediction Center’s Unified Raingauge Database, which does not provide hourly rainfall data. Dixon (2006) used a shorter gulf-surge climatology (1993–2004) and again only acquired daily rainfall data. Becker and Berbery (2008) studied the influence of gulf surges on the diurnal rainfall cycle in southern Arizona by using North American Regional Reanalysis data and determined that maximum daily precipitation totals occurred 3 h earlier during gulf-surge events when compared with nonsurge events. This analysis, however, was limited to the 2004 monsoon season. Therefore, there is a lack of literature that spans multiple monsoon seasons regarding the influence of gulf surges on the diurnal rainfall cycle in this region.

Another relatively unexplored aspect of the gulf-surge phenomenon is the geographical variation of the surge influence on precipitation in Arizona. To the southeast of Arizona, in northwest Mexico, lie the Sierra Madre Occidental and its surrounding lowlands (Fig. 1). This area is in the core of the North American monsoon region, receives the most intense rainfall of any area influenced by the monsoon, and thus serves as an important study area for projects such as the North American Monsoon Experiment (Becker and Berbery 2008). The geographical variation of gulf-surge influences on precipitation in this important region is currently not well understood because of the lack of a reliable climatological database covering the Sierra Madre Occidental (Douglas and Leal 2003). This prevents this study area from being broken up into smaller geographical units (Douglas and Leal 2003). However, the high gauge density and long periods of record present in the National Weather Service Cooperative Observer Program (COOP) network make it possible for these subtle geographic variations to be analyzed in Arizona. Assuming that the mechanisms for thunderstorm development are similar in these two regions, the spatial variability observed in Arizona (e.g., the strength of surge influence may be dependent on elevation) could be similar to the variations around the Sierra Madre Occidental.

Consequently, there are two goals of this paper. First, the regions of Arizona that are most dependent on gulf surges for precipitation will be determined through the analysis of a dense network of rain gauges, all of which had 20–52 monsoon seasons of data. Second, the author will determine the influence of gulf surges on the prominent diurnal rainfall cycle in the Phoenix metropolitan area and its surroundings, which (unlike the rest of Arizona) experience a strong nocturnal precipitation maximum. Knowledge of the locations most affected by gulf surges and the surge effects on diurnal rainfall variability may aid in forecasting thunderstorm development (assuming that gulf surges can be accurately forecast) and highlight the potential catalysts for nocturnal storm development over the Phoenix area.

2. Data and methods

a. Climatology of gulf surges

Many methods have been developed to identify gulf surges (Fuller and Stensrud 2000; Douglas and Leal 2003; Higgins et al. 2004). For example, Douglas and Leal (2003) used radiosonde data from Empalme, Mexico (determining a surge event by an increase in surface-to-850-hPa moisture transport, a decrease in 925-hPa temperature, and an increase in surface pressure), whereas Fuller and Stensrud (2000) as well as Higgins et al. (2004) used surface data from Yuma and Tucson to identify surges (subjectively basing surge onset on rapid increases in surface moisture and southerly wind direction). For this study, the author used the Assessing Low-Level Atmospheric Moisture using Soundings (ALARMS) method developed by Dixon (2005), which identifies a surge event as the third day in any four-day period such that the 850-hPa dewpoint was at least 4°C higher for both of the last two days than for both of the first two days for either the 1200 or the 0000 UTC soundings.

As shown by Dixon (2005), the ALARMS method improves on the alternative gulf-surge detection methods described above. Note that the validation of the ALARMS method by Dixon (2005) was based on surge detection at any one of four radiosonde sites—Empalme, Mexico, and Tucson, Yuma, and Phoenix, Arizona—whereas only data from Tucson were used in this study. Although it is generally accepted that Yuma is the best location for detecting incoming gulf surges, (Fuller and Stensrud 2000; Higgins et al. 2004; Dixon 2006) the lack of missing data (Dixon 2006) and the long period of record for the Tucson radiosonde site make it an adequate location for the purposes of this study. Dixon (2006) suggested that some surges detectable at Yuma through the ALARMS method may not be reflected in the Tucson data; however, Dixon (2006) made no indication of a higher occurrence of incorrectly identified surges, or “false alarms,” at the latter location.

Radiosonde data (0000 and 1200 UTC) from Tucson for the months of July and August and the years 1957–2008 were obtained from the National Oceanic and Atmospheric Administration Earth System Research Laboratory radiosonde database. There were 2.4% and 2.3% missing data for 850-hPa dewpoints for the 0000 and 1200 UTC soundings, respectively. The date of surge onset was identified through the ALARMS method described above; however, surge duration was determined using a method similar to Higgins et al. (2004), who subjectively identified surge onset largely based on rapid increases in surface dewpoint. The end of a surge event was determined by Higgins et al. (2004) as the last day that the surface dewpoint was greater than or equal to the climatological mean for July and August. For 1977–2001, Higgins et al. (2004) identified 111 surges at Tucson and a total of 432 surge days (days occurring through the duration of a surge event), yielding an average surge duration of approximately 4 days. In addition, Higgins et al. (2004) found the average number of surge days per July and August to be 18 days, with a standard deviation of 6 days. Therefore, it seems plausible that numerous surge events identified by Higgins et al. (2004) lasted 5 days or more. Similar to Higgins et al. (2004), the end of a surge was defined as the first day for which both of the soundings showed an 850-hPa dewpoint that was less than the climatological mean for each respective sounding. Additionally, in accordance with the results from Higgins et al. (2004), the maximum surge duration was limited to 5 days. Therefore, a climatology of surge and nonsurge days was created for every July and August for the period 1957–2008 based on surges detected at Tucson by the ALARMS method.

b. Precipitation and gulf surges

Daily rainfall data were obtained for July and August for every station in the Arizona COOP network (available online at http://www.ncdc.noaa.gov/oa/ncdc.html). Only stations recording precipitation for 30 or more years between 1957 and 2008 were considered (Fig. 1). For July and August, the maximum percentage of missing data from these 156 stations was 18.0% and the mean and median among the stations were 4.4% and 3.3%, respectively. Precipitation events from the 156 stations were stratified into presurge and postsurge categories. In accordance with the ALARMS method, the presurge days were the first two days of the four-day period used to identify a given surge and the postsurge days were the last two days of this period. As daily rainfall totals for an arid region such as Arizona are extremely nonnormal because of the high number of days without rain, the statistical significance of the precipitation differences between presurge days and postsurge days needed to be evaluated with caution. A dual method for such an evaluation was employed.

First, a chi-square goodness-of-fit test (Moore 1977) established if there was a significant difference in the number of precipitation events (days on which rainfall occurred) between presurge and postsurge populations. If there was truly no difference between the populations, then it would be expected that the same number of events occurred in each category. Therefore, under the null hypothesis, the number of events should follow a binomial distribution with equal probabilities of an event being in the presurge or postsurge categories. Second, the nonparametric Mood median test (Freidlin and Gastwirth 2000) was used to determine significant differences in the medians of precipitation totals for the presurge and postsurge populations. The results of these tests are given in section 3b.

c. Diurnal rainfall cycle

A record of hourly rainfall data for July and August from 1980 to 2008 was obtained for 63 stations in and around the Phoenix metropolitan area (Fig. 2) from the Flood Control District of Maricopa County (S. Waters 2008, personal communication). Each station had at least 20 years of data. A major quality-control issue was the lack of discrimination between missing data and values of zero prior to 1988. The maximum percentage of missing data from 1988 to 2008 among all hours and stations was 6.1%, and the median was 1.1%.

For each of the 63 stations, the hourly data were stratified by surge and nonsurge days and the total precipitation and precipitation frequency (total nonzero values) were calculated for each hour. Following the methods discussed in Balling and Brazel (1987), harmonic analysis was employed to model the diurnal cycle of precipitation totals and frequencies for surge and nonsurge days. Note that the hourly data could not be stratified by presurge and postsurge days because a sufficient sample size is needed to perform harmonic analysis, and the data of many stations stratified in this manner did not meet the subjective criterion of only one hour recording a total of zero precipitation events. Stratifying by surge and nonsurge days, however, allowed for a sufficient sample size of precipitation events for the majority of stations. Note that Balling and Brazel (1987) required every hour to have at least one rainfall event for harmonic analysis to be performed; however, this requirement led to the elimination of the majority of stations in the surge category. Thus, harmonic analysis (Svoma and Balling 2009) was performed on the 24 rainfall totals and frequencies for 63 stations in the nonsurge category and 35 stations in the surge category. Because the major concern of this study is with the influence of gulf surges on the prominent nocturnal precipitation maximum seen in this region (Balling and Brazel 1987), only the first harmonic was evaluated.

3. Results and discussion

a. Gulf-surge climatology characteristics

As expected, each of the 37 July and August surge events identified by Dixon (2006) at Tucson for 1993–2004 was also identified in this study. For the 52 monsoon seasons considered in this study, 151 individual surge events were identified along with 538 surge days, yielding an average surge duration of 3.6 days. Through simple linear regression, no significant trend (p > 0.10) was identified in the number of surge events or surge days per monsoon season. It is important to note that the error normality and constant error variance assumptions necessary for simple linear regression appeared to be valid for both regression equations. The average (median) number of surge events per July and August was 2.9 (3.0), and the maximum number of events was 7 in 1985 and 2004; no surge events were detected in 1959, 1961, 1965, 1984, 1990, or 1994.

b. Gulf surges and rainfall

Each of the 156 COOP stations had more precipitation events during postsurge times. Only two stations had more total rainfall during the presurge periods, and these differences were miniscule. The mean and median of the fractions of total rainfall and total rain events occurring during postsurge periods for the 156 stations indicate a high dependence of rainfall in Arizona on gulf surges (Tables 1 and 2). The results from the Mood median test and the chi-square goodness-of-fit test indicate that virtually all of the presurge and postsurge differences in rainfall statistics were significant (p < 0.05; Tables 1 and 2). Only four stations showed insignificant differences at the 0.05 level in both the number of rainfall events (chi-square goodness-of-fit test) and total rainfall (Mood median test) between presurge and postsurge cases. The insignificant results for these four stations were likely in response to small total event counts for presurge and postsurge cases combined. These four stations were ranked among the eight lowest stations in terms of total precipitation events recorded, and all had a majority of events occur during postsurge periods.

The results above suggest that, over Arizona as a whole, gulf surges significantly influence precipitation because significantly more rainfall and rain events occurred during postsurge periods than during presurge periods for all but four stations. The spatial variability of this influence can be seen in Fig. 1. The general pattern displayed is that areas that received high amounts of precipitation tended to have lower fractions of precipitation events occur during postsurge periods (locations with darker dots inside smaller circles) and vice versa. For example, the 39 wettest stations had an average of 65% of rain events occur during the postsurge period, whereas an average of 74% of rain events occurred during the postsurge period for the 39 driest stations. This pattern was particularly apparent in the wettest parts of the state during the monsoon season, eastern and southeastern Arizona, as well as the dry central and southwestern deserts, including the Phoenix metropolitan area. This pattern is likely in response to the higher frequency of rain events during nonsurge periods at the wetter stations.

c. Diurnal rainfall cycle

Balling and Brazel (1987) defined a nocturnal rainfall event during the Arizona monsoon season as rainfall occurring during 2000 and 0800 LST. This study used a denser and updated precipitation database for Phoenix and verified the chief finding of Balling and Brazel (1987), which was that nocturnal rainfall events are dominant in the lower elevations around the Phoenix metropolitan area (Fig. 2). Of the 63 stations in central Arizona that were analyzed, 45 had a higher percentage of nocturnal rainfall events during surge periods than during nonsurge periods, suggesting a higher tendency for nocturnal rainfall during gulf-surge events (Fig. 2). This nocturnal maximum was captured well by the first harmonic wave (Fig. 3). Stations with two maxima shown in Fig. 3 are locations for which harmonic analysis was run for both surge and nonsurge periods. Analysis of the first harmonic indicated that there was a tendency for rainfall events to occur later during surge periods, with 30 of these 35 stations having a later maximum of rainfall frequencies during surge periods.

Of the five locations with later maximum rainfall frequencies during nonsurge periods, two had miniscule differences and were located north and south of Phoenix. The three greatest differences were observed at adjacent stations northwest of the valley (Fig. 3). Upon examination of these three stations, it was apparent that the afternoon frequency maximum was largely a result of several light daytime precipitation events during surge periods. These stations were particularly sensitive to these light events because they had relatively small total counts of rain events. Therefore, it is likely that this latter nonsurge maximum is due to influential random variations associated with the smaller sample sizes of these adjacent stations and is not due to any physical mechanism.

These differences in percentages of nocturnal rainfall and first harmonic maxima between surge and nonsurge periods should be taken with caution, however, because spatial autocorrelation is expected among the stations. In other words, considering the close proximity of the stations to one another, we would expect similar surge and nonsurge differences at adjacent stations. This lack of independence in the data population makes it difficult to assess the statistical significance of these differences between surges and nonsurges.

The later maximum for rainfall frequency during surge events is at odds with findings by Becker and Berbery (2008), who found precipitation in southern Arizona to occur earlier during surge events. This may be the case for two reasons. First, this study used a much longer period of record. Second, this study was exclusively for the Phoenix area, which has a unique diurnal precipitation pattern with a strong nocturnal maximum, as mentioned above.

Significance of the first harmonic suggests that the hourly precipitation pattern is indicative of a diurnal cycle. For nonsurge periods, all of the stations had statistically significant (p < 0.05) variance explained by the first harmonic, and the median explained variance among all the stations was a highly significant 65%. Statistical significance was established by taking the square root of the explained variance and assessing the root as a Pearson product-moment correlation coefficient (Rodgers and Nicewander 1988) with 22 degrees of freedom. For surge periods, only seven stations had insignificant explained variance by the first harmonic, and the median was a significant 41% (p < 0.05). The standardized amplitude of the first harmonic is indicative of the importance of a diurnal cycle. For example, a standardized amplitude of 0.10 implies that the probability of occurrence in the peak period is 1.20 times the mean value (Svoma and Balling 2009). The median standardized amplitude of the first harmonic wave was also greater for nonsurge periods (0.32) than for surge periods (0.23). These results suggest that during nonsurge periods temporal rainfall patterns followed a more noticeably defined diurnal cycle than during surge periods; however, as mentioned above, spatial autocorrelation makes the significance of these surge and nonsurge differences difficult to assess.

A detailed study of multiple nocturnal thunderstorm events over the Phoenix area would be essential for determining the most important of the four mechanisms discussed in section 1 for the nocturnal rainfall maximum observed in this area. However, this study illustrates that periods with higher nocturnal precipitation (surge periods) coincide with periods of increased thunderstorm activity in the higher terrain east and northeast of Phoenix (Figs. 1 –3). This suggests that dissipating thunderstorms east and northeast of Phoenix are mechanisms for nocturnal thunderstorm activity in this area. The catalyst provided from these thunderstorms may be the destabilization of the midtroposphere due to evaporatively cooled air and/or cold outflow boundaries, as suggested by McCollum et al. (1995) in a case study of a nocturnal mesoscale convective system over Phoenix.

4. Conclusions

Gulf-surge events, which are associated with the influx of moisture from the Gulf of California into Arizona during the North American monsoon, account for a large portion of annual precipitation occurring in the state (Berbery and Fox-Rabinovitz 2003; Higgins et al. 2004; Becker and Berbery 2008). Using a dense network of rain gauge data, all of which had no fewer than 20 monsoon seasons of data, and the ALARMS method (Dixon 2005) for identifying gulf surges, this study determined 1) the regions of Arizona that are most dependent on gulf surges for precipitation and 2) the influence of gulf surges on the prominent diurnal rainfall cycle in the Phoenix metropolitan area and its surroundings.

Monsoonal precipitation in Arizona was significantly dependent on gulf surges. The strength of this dependence was the largest for areas that receive relatively little precipitation (the deserts of central and southwestern Arizona, including the Phoenix area). This signal was the weakest in the wetter eastern and southeastern portions of the state. Assuming that similar mechanisms for thunderstorm development are present over the Sierra Madre Occidental and its surrounding lowlands (which lies in the core of the North American monsoon region where this spatial variability is not well documented), a similar geographical pattern may be present over this region as well. Additionally, this study suggests that the prominent diurnal cycle and strong nocturnal maximum for rainfall in the Phoenix metropolitan area are evident during both surge and nonsurge periods of the monsoon season. However, this nocturnal maximum was more apparent during surge periods. In conjunction with the higher frequency of precipitation during surge events in the higher terrain to the east and northeast of Phoenix, this suggests that dissipating afternoon thunderstorms are catalysts for the nocturnal maximum.

In short, this study highlights the importance of gulf surges for monsoonal rainfall in Arizona. Assuming that one can predict the occurrence of gulf surges, the implications of these findings should be valuable to summer precipitation forecasting in Arizona for two reasons. One, gulf surges indicate much higher likelihoods for precipitation over Arizona, especially the drier central and southwestern portions of the state. Two, there is a stronger tendency for nocturnal rainfall over the Phoenix metropolitan area during surge events.

Acknowledgments

The author thanks Steve Waters of the Flood Control District of Maricopa County for supplying the hourly precipitation dataset and Dr. Randall Cerveny of the School of Geographical Sciences and Urban Planning at Arizona State University for his guidance.

REFERENCES

  • Anderson, B. T., , Roads J. O. , , and Chen S. C. , 2000: Large-scale forcing of summertime monsoon surges over the Gulf of California and the southwestern United States. J. Geophys. Res., 105 , 2445524467.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Balling R. C. Jr., , , and Brazel S. W. , 1987: Diurnal variations in Arizona monsoon precipitation frequencies. Mon. Wea. Rev., 115 , 342346.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Becker, E. J., , and Berbery E. H. , 2008: The diurnal cycle of precipitation over the North American monsoon region during the NAME 2004 field campaign. J. Climate, 21 , 771787.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Berbery, E. H., , and Fox-Rabinovitz M. S. , 2003: Multiscale diagnosis of the North American monsoon system using a variable-resolution GCM. J. Climate, 16 , 19291947.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Dixon, P. G., 2005: Using sounding data to detect gulf surges during the North American monsoon. Mon. Wea. Rev., 133 , 30473052.

  • Dixon, P. G., 2006: An ALARMS climatology of Gulf of California moisture surges and the associated precipitation, 1993–2004. Phys. Geogr., 27 , 86108.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Douglas, M. W., , and Leal J. C. , 2003: Summertime surges over the Gulf of California: Aspects of their climatology, mean structure, and evolution from radiosonde, NCEP reanalysis, and rainfall data. Wea. Forecasting, 18 , 5574.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Douglas, M. W., , Maddox R. A. , , Howard K. , , and Reyes S. , 1993: The Mexican monsoon. J. Climate, 6 , 16651677.

  • Freidlin, B., , and Gastwirth J. L. , 2000: Should the median test be retired from general use? Amer. Stat., 54 , 161164.

  • Fuller, R. D., , and Stensrud D. J. , 2000: The relationship between tropical easterly waves and surges over the Gulf of California during the North American monsoon. Mon. Wea. Rev., 128 , 29832989.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Hales, J. E., 1977: Relationship of convective cooling to nocturnal thunderstorms at Phoenix. Mon. Wea. Rev., 105 , 16091613.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Higgins, R. W., , and Shi W. , 2005: Relationships between Gulf of California moisture surges and tropical cyclones in the eastern Pacific basin. J. Climate, 18 , 46014620.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Higgins, R. W., , Shi W. , , and Hain C. , 2004: Relationships between Gulf of California moisture surges and precipitation in the southwestern United States. J. Climate, 17 , 29832997.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • King, T. S., , and Balling R. C. , 1994: Diurnal variations in Arizona monsoon lightning data. Mon. Wea. Rev., 122 , 16591664.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Kuwagata, T., , Numaguti A. , , and Endo N. , 2001: Diurnal variation of water vapor over the central Tibetan Plateau during summer. J. Meteor. Soc. Japan, 79 , 401418.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • McCollum, D. M., , Maddox R. A. , , and Howard K. W. , 1995: Case study of a severe mesoscale convective system in central Arizona. Wea. Forecasting, 10 , 643665.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Moore, D. S., 1977: Generalized inverses, Walds method, and construction of chi-squared tests of fit. J. Amer. Stat. Assoc., 72 , 131137.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Reiter, E. R., , and Tang M. C. , 1984: Plateau effects on diurnal circulation patterns. Mon. Wea. Rev., 112 , 638651.

  • Rodgers, J. L., , and Nicewander W. A. , 1988: Thirteen ways to look at the correlation coefficient. Amer. Stat., 42 , 5966.

  • Svoma, B. M., , and Balling R. C. Jr., 2009: An anthropogenic signal in Phoenix, Arizona winter precipitation. Theor. Appl. Climatol., 315321.

    • Search Google Scholar
    • Export Citation
  • Wallace, J. M., 1975: Diurnal variations in precipitation and thunderstorm frequency over the conterminous United States. Mon. Wea. Rev., 103 , 406419.

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

The 156 COOP stations examined. Shown here are the average annual precipitation for July and August (shaded dots) and the fraction of the total presurge and postsurge precipitation events that occurred during the postsurge periods (open circles). The rough location of the Sierra Madre Occidental is shown in the inset as SMO.

Citation: Weather and Forecasting 25, 1; 10.1175/2009WAF2222299.1

Fig. 2.
Fig. 2.

Locations of the 63 hourly rain gauges analyzed around the Phoenix metropolitan area (dots). The percentage of nocturnal rain events (events occurring between 2000 and 0800 LST) during surge periods are represented by open-circle diameter. White (black) dots represent positive (negative) values, which indicate greater (lesser) percentages of nocturnal rainfall events during surge periods.

Citation: Weather and Forecasting 25, 1; 10.1175/2009WAF2222299.1

Fig. 3.
Fig. 3.

Time of maximum for stations for which the first harmonic was calculated, portrayed using “wind direction” symbols with the usual meteorological definition. Flags with open (filled) circles indicate time of maxima for surge (nonsurge) periods. “Northerly” flags indicate 0000 LST, “easterly” flags indicate 0600 LST, etc.

Citation: Weather and Forecasting 25, 1; 10.1175/2009WAF2222299.1

Table 1.

Descriptive statistics for the 156 COOP stations examined in terms of the fraction of the total rainfall that occurred after surge onset. The p values (significance levels) are with respect to chi-square test statistics with 1 degree of freedom.

Table 1.
Table 2.

As in Table 1, but for the fraction of the total rain events that occurred after surge onset.

Table 2.
Save