Introduction
It is estimated that by 2025, 80% of the world's population will live in cities (UNFPA 1999). Urban areas modify boundary layer processes through the creation of an urban heat island (UHI). In cities, natural land surfaces are replaced by artificial surfaces that have different thermal properties (e.g., heat capacity and thermal inertia). Such surfaces are typically more capable of storing solar energy and converting it to sensible heat. Other contributing factors to the onset of the UHI may be attributed to differences in surface albedo and anthropogenic heat release in the urban area. As sensible heat is transferred to the air, the temperature of the air in urban areas tends to be 2°–10°C higher than surrounding nonurban areas. As early as the late nineteenth century, European scientists detected the presence of UHI in Paris, Berlin, Vienna, and London (Landsberg 1956).
In the past 30 years, several observational and climatological studies have theorized that the UHI can have a significant influence on mesoscale circulations and resulting convection. Early investigations (Changnon 1968; Landsberg 1970; Huff and Changnon 1972a,b, 1973) found evidence of warm-season rainfall increases of 9%–17% over and downwind of major urban cities. The Metropolitan Meteorological Experiment (METROMEX) was an extensive study that took place in the 1970s in the United States (Changnon et al. 1977; Huff 1986) to investigate modification of mesoscale and convective rainfall by major cities. In general, results from METROMEX have shown that urban effects lead to increased precipitation during the summer months. Increased precipitation was typically observed within, and 50–75 km downwind of, the city, reflecting increases of 5%–25% over background values (Sanderson and Gorski 1978; Huff and Vogel 1978; Braham and Dungey 1978; Changnon 1979, 1981; Braham et al. 1981; Changnon et al. 1991). Using a numerical model, Hjelmfelt (1982) simulated the UHI of St. Louis and found positive vertical velocities downwind of the city. He suggested that the urban enhanced surface roughness convergence effect and the downwind shifting or enhancement of the UHI circulation by the synoptic flow were the cause. METROMEX results also suggested that areal extent and magnitude of urban and downwind precipitation anomalies were related to size of the urban area (Changnon 1992).
More recent studies have continued to validate and to extend the findings from pre- and post-METROMEX investigations. Balling and Brazel (1987) observed more-frequent late-afternoon storms in Phoenix, Arizona, during recent years of explosive population growth. Analysis by Bornstein and LeRoy (1990) found that New York City affects both summer daytime thunderstorm formation and movement. They illustrated that radar echo maxima were produced on the lateral edges and downwind of the city. Jauregui and Romales (1996) observed that the daytime heat island seemed to be correlated with intensification of rain showers during the wet season (May–October) in Mexico City. They also presented an analysis of historical records showing that frequency of intense rain showers has increased in recent decades in correlation with the growth of the city. Selover (1997) found similar results for moving summer convective storms over Phoenix. Bornstein and Lin (2000) examined data from an Atlanta, Georgia, mesonetwork to show that the UHI induced a convergence zone that initiated storms during the summer of 1996. Thielen et al. (2000) used a meso-gamma-scale model to address the influence of urban surfaces on the development of convective precipitation. The results showed that sensible heat fluxes and enhanced roughness due to the UHI can have considerable influence on convective rainfall. Their results confirmed observations from METROMEX and other studies that the UHI enhances rainfall production over and downwind of the urban area. In addition, the model simulations suggested that stronger heat islands tended to produce more localized effects on rainfall and weaker heat islands affected rainfall at some distance downwind of the heat island. Baik et al. (2001) used a mesoscale cloud model to study dry and moist convection forced by a UHI. They found that a downwind updraft cell's intensity was related to UHI intensity or basic-state wind.
The literature indicates that the signature of the “UHI effect” may be resolvable in rainfall patterns over and downwind of metropolitan areas. However, a recent U.S. Weather Research Program panel concluded that more observational and modeling research is needed because results are not considered definitive at this time (Dabberdt et al. 2000). The National Aeronautics and Space Administration (NASA) and other agencies have initiated programs such as Project ATLANTA (Atlanta Land Use Analysis: Temperature and Air Quality) (Quattrochi et al. 1998). Such programs aim to identify and to understand how UHIs impact the environment.
Previous investigations that studied urban impacts on rainfall used primarily rain gauge networks, ground-based radar, or model simulations. Although useful, these studies were limited to specific cities with special observation networks or theoretical model simulations. In addition, effects of topography and other factors (e.g., sea breezes, lake breezes) were often comparable to the urban effects in cities examined such as St. Louis, Missouri (Huff and Vogel 1978). Herein, a novel approach is introduced to correlate urbanization and rainfall modification. Satellite-based analysis of rainfall modification near urban areas is demonstrated as a viable approach to assessing this problem on a global scale and over longer periods of time.
The Tropical Rainfall Measuring Mission (TRMM) is a joint NASA–National Space Development Agency of Japan mission to study tropical rainfall and its implications for climate. Three years of mean monthly rainfall rates derived from the TRMM precipitation radar (PR) are employed. Analysis of PR data enables identification of rainfall patterns around major metropolitan areas of Atlanta; Montgomery, Alabama; Nashville, Tennessee; and San Antonio, Waco, Austin, and Dallas, Texas. Section 3 will discuss the TRMM data and research methodology employed. The study seeks to identify and to quantify urban modification of rainfall using data from the PR. To be more specific, its objectives are
to validate and to extend ground-based observations of rainfall patterns in major urban areas using satellite rain estimates,
to quantify the impact of urban areas on rainfall in and downwind of cities using satellite rain estimates, and
to demonstrate opportunities to observe multiple urban rainfall anomalies over an extensive area (38°N–38°S) using satellite data.
The study presents a potentially useful complementary dataset to efforts such as Project ATLANTA and future urban–environmental impact studies. The intent of this paper is to demonstrate the ability of space-based radar measurements to identify urban rainfall modification and to corroborate previous findings in this area. The hypothesized factors for rainfall modification by cities are discussed in section 2. Section 2 also provides background on the research strategy for defining the coordinate systems, cities, and time periods of the analysis. Section 3 describes the methodology for obtaining mean monthly rainfall rates from the TRMM PR. Results of contour analyses and statistical calculations are presented in section 4. Section 5 provides a summary and concluding remarks.
Background and research strategy
Urban factors that affect rainfall
Previous research has indicated that urban-induced changes in natural precipitation are most likely due to one or a combination of the following four causes (Changnon et al. 1976; Ochs and Semonin 1979): 1) atmospheric destabilization through enhancement, creation, or displacement of a mesoscale circulation; 2) increased low-level convergence due to surface roughness; 3) modification of microphysical and dynamic processes by the addition of condensation nuclei; or 4) modification of low-level atmospheric moisture content by additions from urban industrial sources. Bornstein and Lin (2000) more recently have suggested that large structures such as aggregations of buildings could act to create a bifurcation zone that steers storms around cities.
An equilibrium surface temperature is required for Eq. (1) to balance. At the surface, if no heat storage is permitted, differential heating results from horizontal gradients in one or more of the terms in Eq. (1). Spatial gradients in this equilibrium temperature in conjunction with the overlying thermodynamic and moisture stratification will dominate the upward or downward flux of heat for thermally forced systems, which results in horizontal temperature gradients required to drive a mesoscale circulation. In the case of the UHI, the difference in surface properties of urban and rural areas leads to the differences in the thermal fluxes in Eq. (1).
Vukovich and Dunn (1978) used a three-dimensional primitive equation model to show that heat island intensity and boundary layer stability have dominant roles in the development of heat island circulations. In addition, Huff and Vogel (1978) found that the urban circulation is primarily enhanced by the increased sensible heat fluxes and surface roughness of the urban area. Hjelmfelt (1982) noted no observation of enhanced glaciation in urban clouds from microphysical changes during METROMEX. Recent research by Rosenfeld (1999) even suggests that increased aerosol amounts may reduce precipitation potential of clouds. In terms of enhancement of rainfall by moisture from industrial sources, Ochs (1975) reported that surface temperature distribution was more important than surface humidity pattern in determining the location of initial cumulus activity in his two-dimensional model simulations investigating the impact of urban surfaces. Orville et al. (1981) used simulations to show that sensible heat rather than moisture (e.g., latent heat) from nuclear power parks had a larger effect on convective clouds. Most studies suggest that dynamic forcing (e.g., heat island destabilization and surface roughness) are more significant to urban rainfall modification than microphysical or moisture enhancement. More definitive research is needed in this area, however.
Definition of “working” control coordinate system
To investigate the capabilities of satellite-based measurements for identifying urban effects on rainfall, a working hypothesis was established, similar in philosophy to Huff and Changnon (1972a). In their framework, hypothesized areas of urban effect and no effect on a climatological timescale were determined. Their study identified the most frequent lower-tropospheric wind flow for each city and defined the hypothesized “downwind affected region” and upwind control regions. Our working hypothesis is a variation of this approach (Fig. 1):
Areas within a 25-km radius of the city (e.g., the central urban area) will exhibit some level of enhanced precipitation due to the UHI effects.
Areas within 25–75 km downwind of the central urban area and within a 125° sector will exhibit the maximum impact area (MIA) of UHI effects.
Areas within 25–75 km upwind of the central urban area are defined as the “upwind control area (UCA).”
Areas within approximately 50 km2 orthogonal to the mean wind vector are considered to be minimal-to-no-impact regions.
The gray arrow in Fig. 1 represents a mean wind vector at 700 hPa from the direction 270° (i.e., “westerly”) and is the horizontal reference axis (HRA) that determines the orientation of the control coordinate system. The 700-hPa level was chosen arbitrarily as a representative level for the mean steering flow for convective storms and is supported by previous work in the literature (Hagemeyer 1991). The National Centers for Environmental Prediction–National Center for Atmospheric Research (NCEP–NCAR) reanalysis dataset (Kalnay et al. 1996) and published work by Hagemeyer (1991) were used to determine the mean warm-season “prevailing” flow at 700 hPa for the selected cities. The mean May–September 700-hPa geopotential height from 1979 to 1998 was analyzed and summarized in Table 1. For each city, the HRA is oriented according to the mean prevailing wind direction. The 125° sector accounts for the “steering” directions representing means that include values greater or less than the mean value; therefore, the MIA accounts for the spread of values that encompass the mean (e.g., the deviation).
Rationale for selected cities and time period of study
A potential shortcoming of any study that attempts to link rainfall modification with urban areas is the difficulty of separating topographic and other effects (e.g., sea-breeze circulations, river-breeze circulations) from urban effects. An early investigation by Huff and Changnon (1973) examined rainfall modification in eight urban areas. Two of the cities—Cleveland, Ohio, and Chicago, Illinois—bordered major Great Lakes and possibly experienced lake-breeze circulations, and two—New Orleans, Louisiana, and Houston, Texas—were coastal and likely experienced sea-breeze forcing regularly. In addition, Washington, District of Columbia, is just east of significant topography and west of the Chesapeake Bay. St. Louis investigations, the focus of much of the METROMEX work, also had to confront issues related to topography. Huff and Vogel (1978) demonstrated some success in separating urban and topographic effects on summer rainfall distribution around St. Louis following METROMEX. The Mississippi River and its river breeze could also impact mesoscale circulations near St. Louis. These factors suggest caution when considering urban circulations against other mesoscale-induced circulations.
The cities chosen for this study (Table 1) are Atlanta, Montgomery, Dallas, Waco, Austin, and San Antonio. Nashville was also examined; however, observational restraints of the satellite (section 3) and data contamination (section 4) left this site with incomplete data. Each city was selected because it was 1) located between TRMM's 38°N and 38°S latitudinal boundary, 2) relatively flat, 3) not located near major topography or major water–land boundaries, and 4) representative of an urban area with clearly distinguishable urban and rural zones.
The study also focused only on the warm-season months (May–September) of 1998–2000 because the 3-yr period reflects the availability of TRMM data at the time of writing. Changnon et al. (1991) found some evidence that St. Louis could also alter precipitation in the autumn, winter, and spring; however, the overwhelming consensus from METROMEX and other efforts is that urban effects are most pronounced during warm-season months (Huff and Changnon 1972a; Changnon et al. 1991; Jauregui and Romales 1996). The most likely reason for pronounced urban effects during the warm season is smaller large-scale synoptic forcing. Advection associated with relatively strong synoptic flow tends to eliminate the thermal differentiation between rural and urban areas (Pielke and Segal 1986), and synoptic forcing such as frontal systems tends to mask mesoscale circulations during the cool season. During the warm season, the UHI-induced mesoscale circulation is more dominant and can significantly alter boundary layer processes.
Data and methodology
One of the novelties of this study is the application of satellite data to the problem of rainfall modification by urban areas. Although the dataset is not sufficient to establish true climatological behavior, the 3-yr period does allow for the investigation of anomalies in rainfall associated with mesoscale circulations. The appearance of distinct “mesoscale signatures” in the TRMM rainfall-rate analysis suggested that the dataset might be useful for identifying urban mesoscale rainfall signatures. The emphasis here is not to duplicate the methodology of historical climatological studies on urban rainfall modification but to establish the capability of satellite-based rainfall estimates to identify urban rainfall anomalies in multiple cities. The data examined provide evidence of mesoscale signatures in rainfall distributions and first steps in the use of satellite-based rainfall estimates in urban modification studies.
This study utilizes a spaceborne radar dataset from the TRMM satellite, which was launched in November of 1997 as a joint U.S.–Japanese mission to advance understanding of the global energy and water cycle by providing distributions of rainfall and latent heating over the global Tropics. The TRMM PR operates at a frequency of 13.8 GHz and can achieve quantitative rainfall estimation over land and ocean (Fig. 2). The horizontal resolution of 4.3 km at nadir and about 5 km at the scan edge allows the TRMM PR to observe small convective cells in addition to larger systems. TRMM is in a low-inclination (35°), low-altitude orbit and is particularly suited for capturing rainfall events at temporal scales of 1 month or greater. Because of the non-sun-synchronous orbit strategy, the satellite precesses through the diurnal cycle roughly every 23 days. For this reason, it is unlikely that our results reflect any biases from diurnal forcing.
The primary sampling issues for this study are related to sampling area and occurrence of rainfall. Figure 3 provides results from quantifying how often the TRMM PR samples a given area in a 15-month period, as a function of latitude and longitude. The rectangular regions represent the approximate study areas addressed herein. Using a gridded set of global 0.5°-resolution data for each orbit and a calculation based on Bell and Reid (1993), an analysis of the number of rainfall estimates in a 0.5° × 0.5° cell for any given hour was made. This number was multiplied by the quantity (24 h × 30 days × 15 months) to retrieve the number of samples at a given latitude and longitude over a 15-month period. As Bell and Reid (1993) discussed, Figure 3 illustrates that observations are more frequent near higher latitudes and become more infrequent toward the equator. It is fortuitous that the two focus areas of this study are within regions of fairly robust samples ranging from 300 to 800 samples. The maximum number of samples is found near the 34°-latitude region (near orbit-crossing regions), and the primary dependence of sampling frequency is latitudinal. As for rain frequency, the rainfall products used in this study are based on an algorithm (to be discussed later) in which monthly statistics in space and time are computed. The version of the product utilized herein calculates rain statistics only when rain is judged to be certain (as opposed to clutter, noise, etc.) in a 0.5° cell. Therefore, we are fairly confident that this study contains a robust number (order of hundreds) of samples in each 0.5°-resolution cell over the 15-month study period.
To assess the accuracy of the spaceborne radar system, we refer to recent literature citing the performance of the PR. Kummerow et al. (2000) reported that comparisons of PR-measured radar reflectivities with those measured by ground-based radar at NASA's Florida ground validation site show good agreements (differences within about 1 dB, on average). Schumacher and Houze (2000) compared the PR rainfall estimates with an S-band ground-based radar in Kwajelin Atoll and also found good agreement. They found that the PR only misses 2.3% of near-surface rainfall relative to the ground-based radar and gauges. Similar calibration and validation studies corroborate these results (Bolen and Chandrasekar 2000; Heymsfield et al. 2000). The validation and calibration results indicate that the PR has been and will be sufficiently stable and accurate to assure quantitative rainfall estimates. In fact, operational agencies are considering the possibility of using the PR as a calibration constant to ground-based radars that are calibrated independently and rarely to the 1-dBZ standards of the PR (Kummerow et al. 2000).
Space–time-averaged PR rainfall products are utilized to investigate rainfall modification by urban effects. The analysis is primarily conducted on mean monthly rainfall rates (mm h−1) at a height of 2.0 km in 0.5° × 0.5° cells. These rainfall rates were calculated as a part of the algorithm described in the following section (Fig. 4). In the diagram, unprocessed receiver counts are converted to calibrated received power with the standard radar equation. Conversion equations are then used to convert PR received power into radar reflectivity factor. This procedure is consistent with radar data processing with ground-based systems.
The final step in the processing involves conversion of reflectivity Z to rainfall rate estimates R using a Z–R relationship. As with ground-based radar systems, the specific Z–R relationship may vary depending on region, rain type, or season. The relationship selected is R = aZb, in which a and b are functions of the rain type, existence of brightbrand, freezing height, storm height, and absolute height. Effects of the difference in raindrop size distribution by rain type, phase state, temperature, and the difference in terminal velocity due to changes in the air density with height are also accounted for. Figure 5 is an example of a mean monthly, 0.5°-resolution, rainfall map for June of 2000 centered over Texas.
For more detailed analysis, the mean rainfall rate value at each grid point was calculated over the 3-yr period (1998–2000) for the months of May, June, July, August, and September. For a given point, a total of 15 mean monthly rainfall-rate values were averaged (3 yr × 5 months of data). This procedure effectively establishes a 15-month analysis of PR rainfall rates for 1998–2000. It is important to remember that monthly rainfall rates at each grid point are aggregates of numerous pixels defined as “rainy” by the algorithm over the 30-day period. The mean values were placed in Cartesian coordinates and contoured.
Results
Overview analysis of southeastern cities
Analysis reveals findings consistent with previous studies that employed ground measurements or numerical models. A Geostationary Operational Environmental Satellite (GOES) IR (channel 2, 3.9 μm) image of the southeastern United States is placed next to the analysis of the 15-month rainfall rates (Fig. 6a), where the arrows indicate dark (e.g., warm) thermal signatures associated with UHI in Atlanta, Montgomery, and Nashville.
Recent work from Project ATLANTA by Bornstein and Lin (2000) discovered that UHIs create thunderstorms in downwind quadrants of the city in the cases examined. Rainfall amounts were calculated from a special network of high-density rain gauge networks placed around Atlanta. Note that the underlying dataset in Fig. 6b is from NASA's Landsat-5 spacecraft and illustrates urban and rural landscapes. The TRMM data (Fig. 6b) indicate a relative maximum in warm-season rainfall rates slightly southeast of the city, consistent with the placement of UHI-induced rainfall anomalies such as those reported by Bornstein and Lin (2000). For the TRMM data in Fig. 6b, values contoured in red (blue) represent values of at least 4.4 mm h−1 (less than or equal to 3.6 mm h−1). The relative maximum in Fig. 6b falls within the MIA (Fig. 1 and Table 1) and the relatively smaller rainfall rates west and north of the city, as hypothesized. This maximum is also consistent with an analysis of rainfall totals collected by the Georgia Automated Environmental Monitoring Network (AEMN; Hoogenboom 1996). Rainfall totals were binned in the same manner as the TRMM PR data from 1998 to 2000. Figure 7b illustrates a broad maximum south-southeast of the city during the May–September period. Though Fig. 7b provides rainfall totals and Fig. 6b provides rainfall rates, the consistency of the relative maxima locations provides encouraging validation for the PR. A close examination of Fig. 7a reveals that areas (particularly southeastern sections) of metropolitan Atlanta are poorly sampled by the AEMN network, thereby leading to possible biases in the data. This point illustrates the need for higher-density networks near cities to validate the satellite estimates.
Montgomery (MGM) is a smaller urban area than Atlanta, but Fig. 6a illustrates that it can still generate a UHI. TRMM data further suggest that Montgomery's UHI exerts an influence on rainfall. Based on the current hypothesis, the maximum impact area should be east of the city and within the 125° sector (Fig. 1 and Table 1). Relatively high rates are found in the MIA while relatively low values are found west and north of the city (Fig. 6b). Evidence also existed that rainfall is maximized in the MIA of Nashville; however, Nashville (BNA) was excluded from the analysis because there were insufficient data to compute values in the UCA because Nashville is near the northern latitudinal extent of TRMM coverage.
A more quantitative assessment of how the various sectors of our hypothesized control coordinate system varied in terms of rainfall rates around Atlanta and Montgomery was made. Table 2 indicates that the mean rainfall rates in the MIA for Atlanta and Montgomery were greater than the UCA by 19.5% and 14.6%, respectively. This number is consistent with findings by Huff and Changnon (1973), who found warm-seasonal rainfall increases of 9%–17% for large cities. The data indicate smaller increases directly over the urban area for Atlanta (7.8%) and Montgomery (9.9%), which suggests a greater enhancement downwind of the city.
Overview analysis of Texas cities
The Texas GOES IR image (Fig. 8a) illustrates the significant UHI signatures associated with urban centers along the Interstate-35 (I-35) corridor (i.e., Dallas, Waco, Austin, and San Antonio) and in Houston. Of interesting, an examination of Fig. 8b indicates that the 15-month rainfall rate analysis from TRMM reveals relative maxima approximately 30–100 km east or northeast of Dallas (DFW), Waco (ACT), Austin (AUS), and San Antonio (SAT). At the same time, relatively minimal rainfall rates are found west of these cities. In fact, the rainfall maxima were strong identifiers of the cities and I-35 corridor before overlaying of navigation markers or underlaying of Landsat-5 data. It is also interesting to observe the two rainfall maxima to the east and west of Galveston Bay [near Houston (HOU)]. It can be argued that at least one of these maxima has UHI influences; however, it is difficult and beyond the scope of this paper to differentiate impacts related to sea-breeze circulation. Houston is not included in the analysis but will be examined in future work.
The mean prevailing steering flows for Dallas, Waco, and San Antonio were 225°, 210°, and 198°, respectively (Table 1). Using the control coordinate system, it is found that the relative maxima in Fig. 8b are all located in the downwind MIA. Results indicate that the mean rain rate in the MIA (urban center) for Dallas was 32% (24.7%) greater than in the UCA. For Waco, the mean rain rate in the MIA (urban center) was 51.1% (14.7%) greater than in the UCA. Consistent results are found for San Antonio, with the mean rain rate in the MIA (urban center) exhibiting a difference of 25.5% (−27.7%) over the UCA. Though not included in Table 1, the mean rain rate in the MIA (urban center) for Austin was 68% (41%) larger than values in the UCA. Portions of the MIA for Austin and San Antonio overlap, so the focus was put on the larger metropolitan area of San Antonio; however, it is likely that Austin exerts significant influence on UHI-induced rainfall.
To evaluate the significance of the differences in warm-season rainfall rates between the upwind control area and hypothesized effect areas, statistical t tests were applied. The t test gives the probability that the difference between the mean of two groups is caused by chance rather than some forcing or circumstance. It is customary to establish that if this probability is less than 0.05, the difference is significant and not caused by chance. Significance testing indicates the following mean values for all cities in the study:
major impact area versus upwind control, probability = 0.034;
urban area versus upwind control, probability = 0.805.
General summary of results
The current results corroborate early METROMEX findings and later studies from ground observations that suggested a downwind maximum in rainfall relative to major urban cities. For the five cities studied, the average 3-yr, warm-season rainfall rates were 28.4% larger in the downwind “maximum impact area” than in the upwind control area defined. The rates were 5.8% greater over the urban city. This value increases to 14.2% if the negative value for San Antonio is not included. In the minimum impact area to the left (right) of the prevailing wind vector, rainfall rates were 1.1% (10.7%) greater than in the UCA.
The results suggest a definite bias toward greater enhancement in the downwind regions of the urban area with minimal enhancement directly over the city and orthogonal to the prevailing wind vector. This downwind bias is also apparent in analysis of the URR defined in Eq. (2). URR rates for all grid points in the control coordinate system for all five cities in the study (Fig. 9) show that 70% of the values above a reference value of 1.0 (i.e., the threshold for positive anomalies) are found in the downwind maximum impact area. It also reveals that the majority of upwind control points (76%) have URR values less than 1.0, which indicates a downwind bias toward increased rainfall rates. The majority of the points in the minimum impact area fall below URR values of 1.0 also; values over the urban center generally cluster close to 1.0. However, the important finding verifies the tendency toward higher rates downwind of the city. In previous studies (Landsberg 1970; Changnon 1968; Huff and Changnon 1973; Sanderson and Gorski 1978; Changnon et al. 1991; Thielen et al. 2000), it was found that summer precipitation values in and downwind of the city reflected increases of 5%–25% over background values. This is consistent with the range of 5.8%–28.4% in the TRMM rain-rate fields and the analysis of URR values in Fig. 9.
The next question of interest is to determine how far downwind the primary urban-influenced rainfall maxima occur. Huff and Changnon (1972a) found that rainfall within a radius of 50–75 miles of St. Louis was impacted by the city. Thielen et al. (2000) noted that METROMEX investigators reported enhancement over and at a distance of 40 km downwind of St. Louis. Thielen et al. (2000) also reported that rainfall was focused over and 60–80 km downwind of the urban surface in their “urban” model simulation.
To investigate the distance factor using the 3-yr TRMM dataset, we identified the location and distance of the maximum rainfall rate found in the maximum impact area (downwind) of each city (Table 3). In general, the maximum value is greater than the mean value of the upwind control area by a range of 48.5%–116%. Also, the maximum value lies in the MIA at distances ranging from 20 to 60 km from the edge of the urban area (45–85 km from the exact center). Overall, the maximum value is found in the MIA at a mean distance of 39 km from the edge of the urban enter or 64 km from the exact center. Again, these values are consistent with findings from previous investigators. Figure 10 is a schematic summary of the general location and distance of the area near each city that likely exhibited urban-impacted rainfall modification.
Summary and conclusions
The primary goal of this study was to establish that a 3-yr, warm-season analysis of mean rainfall rates from the TRMM PR could be used to identify urban-induced rainfall anomalies. The study also demonstrated the potential capability to study urban rainfall modification on a global scale and over longer time periods.
With the recollection that prevailing wind was determined based on a 19-yr climatic description of geopotential heights, the results validated previous ground-based and modeling studies that identified urban-induced rainfall maxima over and downwind of cities. Using a 15-month (spanning 3 years) analysis of mean rainfall rates, we examined the cities of Atlanta, Montgomery, Dallas, Waco, and San Antonio. We found that the average percentage increase in mean rainfall rate in the hypothesized downwind affected region over the upwind control area was 28.4% with a range of 14.6%–51%. Over the urban area, the average change was smaller (+5.8%) but exhibited a range of −27.7%–24.7%. There was a slight indication that regions orthogonal and to the right of the mean prevailing flow (within 50 km) experienced relatively significant increases in rainfall (10.7%). However, the downwind region exhibited the most significant changes.
We also demonstrated that the maximum rainfall rates found in the maximum impact area exceeded the mean value in the upwind control area by 48%–116%. This maximum value was found at an average distance of 39 km from the edge of the urban center or 64 km from the exact center. The range was 20–60 km downwind of the edge of the urban center. In general, the changes in rainfall and their location relative to the “nonurban” effect regions are consistent with previous work related to METROMEX and other studies. This fact provides confidence that UHI-rainfall effects are real and that satellite rainfall estimates from TRMM can detect them.
The implications of the research presented herein are broad. The establishment of TRMM's ability to identify rainfall anomalies associated with urban areas provides a powerful tool to investigate urban effects due to other world cities between 38°N and 38°S, particularly in areas with sparse ground-based rain measurement systems. The future space-based rainfall measuring missions (e.g., Global Precipitation Measurement) will extend TRMM-like measurements to the midlatitudes, thereby extending our approach to numerous major cities not located in the subtropical and tropical latitudes that TRMM observed. As experimental and real-time weather prediction models continue to approach smaller spatial scales, this research may require mesoscale models to consider urban surfaces and their characteristics in surface–land parameterizations. This point is particularly critical because urban growth continues to infringe upon green space at alarming rates. In addition, the research has implications for policy makers, urban planners, water resource managers, and agriculture professionals who may use an understanding of urban rainfall climate in the design of better drainage systems, planning of land use, or identification of optimal areas for agricultural activity. The study also demonstrates the impact of human development on environmental processes.
Acknowledgments
The authors thank Drs. Ramesh Kakar and Robert Adler for providing support for this research through NASA's TRMM Project. The authors also thank Dr. Dennis Chesters for valuable insight on GOES images. We are grateful to Drs. David Starr and Tom Bell for guidance in this project and to our colleagues who agreed to review this manuscript and provided valuable comments and suggestions.
REFERENCES
Baik, J-J., Y. H. Kim, and H. Y. Chun. 2001. Dry and moist convection forced by an urban heat island. J. Appl. Meteor. 40:1462–1475.
Balling, R. and S. Brazel. 1987. Recent changes in Phoenix summertime diurnal precipitation patterns. Theor. Appl. Climatol. 38:50–54.
Bell, T. and N. Reid. 1993. Detecting the diurnal cycle of rainfall using satellite observations. J. Appl. Meteor. 32:311–322.
Bolen, S. M. and V. Chandresekar. 2000. Quantitative cross validation of space-based and ground-based radar. J. Appl. Meteor. 39:2071–2079.
Bornstein, R. and G. M. LeRoy. 1990. Urban barrier effects on convective and frontal thunderstorms. Extended Abstracts, Fourth Conf. on Mesoscale Processes, Boulder, CO, Amer. Meteor. Soc., 120–121.
Bornstein, R. and Q. Lin. 2000. Urban heat islands and summertime convective thunderstorms in Atlanta: Three case studies. Atmos. Environ. 34:507–516.
Braham, R. R. Jr, and M. Dungey. 1978. A study of urban effects on radar first echoes. J. Appl. Meteor. 17:644–654.
Braham, R. R. Jr,, R. G. Semonin, A. H. Auer, S. A. Changnon Jr., and J. M. Hales. 1981. Summary of urban effects on clouds and rain. METROMEX: A Review and Summary, Meteor. Monogr., No. 40, Amer. Meteor. Soc., 141–152.
Changnon, S. A. Jr, 1968. The La Porte weather anomaly—fact or fiction? Bull. Amer. Meteor. Soc. 49:4–11.
Changnon, S. A. Jr, . 1979. Rainfall changes in summer caused by St. Louis. Science 205:402–404.
Changnon, S. A. Jr, . Ed.,. . 1981. METROMEX: A Review and Summary. Meteor. Monogr., No. 40, Amer. Meteor. Soc., 181 pp.
Changnon, S. A. Jr, . 1992. Inadvertent weather modification in urban areas: Lessons for global climate change. Bull. Amer. Meteor. Soc. 73:619–627.
Changnon, S. A. Jr,, R. G. Semonin, and F. A. Huff. 1976. A hypothesis for urban rainfall anomalies. J. Appl. Meteor. 15:544–560.
Changnon, S. A. Jr,, F. A. Huff, P. T. Schickendanz, and J. L. Vogel. 1977. Summary of METROMEX, Vol. 1: Weather anomalies and impacts. Illinois State Water Survey Bull. 62, 260 pp.
Changnon, S. A. Jr,, R. T. Shealy, and R. W. Scott. 1991. Precipitation changes in fall, winter, and spring caused by St. Louis. J. Appl. Meteor. 30:126–134.
Dabberdt, W. F. Coauthors,. 2000. Forecast issues in the urban zone: Report of the 10th prospectus development team of the U.S. Weather Research Program. Bull. Amer. Meteor. Soc. 81:2047–2064.
Hagemeyer, B. C. 1991. A lower-tropospheric thermodynamic climatology for March through September: Some implications for thunderstorm forecasting. Wea. Forecasting 6:254–270.
Heymsfield, G. M., B. Geerts, and L. Tian. 2000. TRMM precipitation radar reflectivity profiles as compared with high-resolution airborne and ground-based radar measurements. J. Appl. Meteor. 39:2080–2102.
Hjemfelt, M. R. 1982. Numerical simulation of the effects of St. Louis on mesoscale boundary-layer airflow and vertical air motion: Simulations of urban vs non-urban effects. J. Appl. Meteor. 21:1239–1257.
Hoogenboom, G. 1996. The Georgia Automated Environmental Monitoring Network. Preprints, 22d Conf. on Agricultural and Forest Meteorology with Symp. on Fire and Forest Meteorology, Atlanta, GA, Amer. Meteor. Soc., 343–346.
Huff, F. A. 1986. Urban hydrological review. Bull. Amer. Meteor. Soc. 67:703–712.
Huff, F. A. and S. A. Changnon Jr.. 1972a. Climatological assessment of urban effects on precipitation at St. Louis. J. Appl. Meteor. 11:823–842.
Huff, F. A. . 1972b. Climatological assessment of urban effects on precipitation, Part II. Final Report, NSF Grant GA-18781, Illinois State Water Survey, 237 pp.
Huff, F. A. . 1973. Precipitation modification by major urban areas. Bull. Amer. Meteor. Soc. 54:1220–1232.
Huff, F. A. and J. L. Vogel. 1978. Urban, topographic and diurnal effects on rainfall in the St. Louis region. J. Appl. Meteor. 17:565–577.
Jauregui, E. and E. Romales. 1996. Urban effects on convective precipitation in Mexico City. Atmos. Environ. 30:3383–3389.
Kalnay, E. Coauthors,. 1996. The NCEP/NCAR 40-Year Reanalysis Project. Bull. Amer. Meteor. Soc. 77:437–471.
Kummerow, C. Coauthors,. 2000. The status of the Tropical Rainfall Measuring Mission (TRMM) after two years in orbit. J. Appl. Meteor. 39:1965–1982.
Landsberg, H. E. 1956. The climate of towns. Man's Role in Changing the Face of the Earth. University of Chicago Press, 584–603.
Landsberg, H. E. . 1970. Man-made climate changes. Science 170:1265–1274.
National Space Development Agency of Japan and National Aeronautics and Space Administration, 2000. Tropical Rainfall Measuring Mission precipitation radar algorithm instruction manual, version 2.0. TRMM Precipitation Radar Team, 113 pp. [Available from NASA Goddard Space Flight Center, Earth Sciences Directorate, Greenbelt, MD 20771.].
Ochs, H. T. I. I. I, 1975. Modeling of cumulus initiation in METROMEX. J. Appl. Meteor. 14:873–882.
Ochs, H. T. I. I. I, and R. G. Semonin. 1979. Sensitivity of a cloud microphysical model to an urban environment. J. Appl. Meteor. 18:1118–1129.
Orville, H. D., P. A. Eckhoff, J. E. Peak, J. H. Hirsch, and F. J. Kopp. 1981. Numerical simulation of the effects of cooling tower complexes on clouds and severe storms. Atmos. Environ. 15:823–835.
Pielke, R. A. and M. Segal. 1986. Mesoscale circulations forced by differential terrain heating. Mesoscale Meteorology and Forecasting, Peter S. Ray, Ed., Amer. Meteor. Soc., 793 pp.
Quattrochi, D. A. Coauthors,. 1998. Project ATLANTA (ATlanta Land use ANalysis: Temperature and Air quality)—A study of how the urban landscape affects meteorology and air quality through time. Preprints, Second Urban Environment Symp., Albuquerque, NM, Amer. Meteor. Soc., 104–107.
Rosenfeld, D. 1999. TRMM observed first direct evidence of smoke from forest fires inhibiting rainfall. Geophys. Res. Lett. 26:3105–3108.
Sanderson, M. and R. Gorski. 1978. The effect of metropolitan Detroit–Windsor on precipitation. J. Appl. Meteor. 17:423–427.
Schumacher, C. and R. A. Houze Jr.. 2000. Comparison of radar data from the TRMM satellite and Kwajalein oceanic validation site. J. Appl. Meteor. 39:2151–2164.
Selover, N. 1997. Precipitation patterns around an urban desert environment—topographic or urban influences? Association of American Geographers Annual Meeting Abstracts, Fort Worth, TX, AAG.
Thielen, J., W. Wobrock, A. Gadian, P. G. Mestayer, and J-D. Creutin. 2000. The possible influence of urban surfaces on rainfall development: A sensitivity study in 2D in the meso-gamma scale. Atmos. Res. 54:15–39.
UNFPA, 1999. The State of World Population 1999. United Nations Population Fund, United Nations Publications, 76 pp.
Vukovich, F. M. and J. W. Dunn. 1978. A theoretical study of the St. Louis heat island: Some parameter variations. J. Appl. Meteor. 17:1585–1594.
Theoretical coordinate system used to define upwind control, urban, and maximum UHI-rainfall impact area. Gray arrow depicts the mean prevailing wind and defines the reference axis for the coordinate system
Citation: Journal of Applied Meteorology 41, 7; 10.1175/1520-0450(2002)041<0689:RMBMUA>2.0.CO;2
Theoretical coordinate system used to define upwind control, urban, and maximum UHI-rainfall impact area. Gray arrow depicts the mean prevailing wind and defines the reference axis for the coordinate system
Citation: Journal of Applied Meteorology 41, 7; 10.1175/1520-0450(2002)041<0689:RMBMUA>2.0.CO;2
Theoretical coordinate system used to define upwind control, urban, and maximum UHI-rainfall impact area. Gray arrow depicts the mean prevailing wind and defines the reference axis for the coordinate system
Citation: Journal of Applied Meteorology 41, 7; 10.1175/1520-0450(2002)041<0689:RMBMUA>2.0.CO;2
Schematic of TRMM platform scan strategy (following National Space Development Agency of Japan and National Aeronautics and Space Administration 2000)
Citation: Journal of Applied Meteorology 41, 7; 10.1175/1520-0450(2002)041<0689:RMBMUA>2.0.CO;2
Schematic of TRMM platform scan strategy (following National Space Development Agency of Japan and National Aeronautics and Space Administration 2000)
Citation: Journal of Applied Meteorology 41, 7; 10.1175/1520-0450(2002)041<0689:RMBMUA>2.0.CO;2
Schematic of TRMM platform scan strategy (following National Space Development Agency of Japan and National Aeronautics and Space Administration 2000)
Citation: Journal of Applied Meteorology 41, 7; 10.1175/1520-0450(2002)041<0689:RMBMUA>2.0.CO;2
A diagram illustrating the number of rainfall estimates in 0.5°-resolution cells as a function of latitude and longitude for region A (Texas region) and region B (southeastern United States) for a 15-month sample period. Sample density decreases as latitude approaches the equator
Citation: Journal of Applied Meteorology 41, 7; 10.1175/1520-0450(2002)041<0689:RMBMUA>2.0.CO;2
A diagram illustrating the number of rainfall estimates in 0.5°-resolution cells as a function of latitude and longitude for region A (Texas region) and region B (southeastern United States) for a 15-month sample period. Sample density decreases as latitude approaches the equator
Citation: Journal of Applied Meteorology 41, 7; 10.1175/1520-0450(2002)041<0689:RMBMUA>2.0.CO;2
A diagram illustrating the number of rainfall estimates in 0.5°-resolution cells as a function of latitude and longitude for region A (Texas region) and region B (southeastern United States) for a 15-month sample period. Sample density decreases as latitude approaches the equator
Citation: Journal of Applied Meteorology 41, 7; 10.1175/1520-0450(2002)041<0689:RMBMUA>2.0.CO;2
TRMM algorithm flow chart (following National Space Development Agency of Japan and National Aeronautics and Space Administration 2000)
Citation: Journal of Applied Meteorology 41, 7; 10.1175/1520-0450(2002)041<0689:RMBMUA>2.0.CO;2
TRMM algorithm flow chart (following National Space Development Agency of Japan and National Aeronautics and Space Administration 2000)
Citation: Journal of Applied Meteorology 41, 7; 10.1175/1520-0450(2002)041<0689:RMBMUA>2.0.CO;2
TRMM algorithm flow chart (following National Space Development Agency of Japan and National Aeronautics and Space Administration 2000)
Citation: Journal of Applied Meteorology 41, 7; 10.1175/1520-0450(2002)041<0689:RMBMUA>2.0.CO;2
Mean rainfall rates at a height of 2.0 km. The bottom panel illustrates global rainfall rates for Jun 2000. The top panel focuses on a region centered over Texas. Color legend (mm h−1): blue shades (0.0–2.0), dark green shades (2.0–5.0), light green (5.0–8.00), tan and yellow shades (8.0–16.0), and orange and red shades (16.0–100.0)
Citation: Journal of Applied Meteorology 41, 7; 10.1175/1520-0450(2002)041<0689:RMBMUA>2.0.CO;2
Mean rainfall rates at a height of 2.0 km. The bottom panel illustrates global rainfall rates for Jun 2000. The top panel focuses on a region centered over Texas. Color legend (mm h−1): blue shades (0.0–2.0), dark green shades (2.0–5.0), light green (5.0–8.00), tan and yellow shades (8.0–16.0), and orange and red shades (16.0–100.0)
Citation: Journal of Applied Meteorology 41, 7; 10.1175/1520-0450(2002)041<0689:RMBMUA>2.0.CO;2
Mean rainfall rates at a height of 2.0 km. The bottom panel illustrates global rainfall rates for Jun 2000. The top panel focuses on a region centered over Texas. Color legend (mm h−1): blue shades (0.0–2.0), dark green shades (2.0–5.0), light green (5.0–8.00), tan and yellow shades (8.0–16.0), and orange and red shades (16.0–100.0)
Citation: Journal of Applied Meteorology 41, 7; 10.1175/1520-0450(2002)041<0689:RMBMUA>2.0.CO;2
(a) A GOES IR 3.9-μm image of the Southeast. Urban heat islands for Nashville, Montgomery, and Atlanta are indicated as dark warm regions in the circles. (b) A contour plot of the 15-month, warm-season mean rainfall rates at a height of 2.0 km using 0.5°-resolution TRMM PR data. Values in red are greater than or equal to 4.2 mm h−1. Values in blue are less than or equal to 3.6 mm h−1
Citation: Journal of Applied Meteorology 41, 7; 10.1175/1520-0450(2002)041<0689:RMBMUA>2.0.CO;2
(a) A GOES IR 3.9-μm image of the Southeast. Urban heat islands for Nashville, Montgomery, and Atlanta are indicated as dark warm regions in the circles. (b) A contour plot of the 15-month, warm-season mean rainfall rates at a height of 2.0 km using 0.5°-resolution TRMM PR data. Values in red are greater than or equal to 4.2 mm h−1. Values in blue are less than or equal to 3.6 mm h−1
Citation: Journal of Applied Meteorology 41, 7; 10.1175/1520-0450(2002)041<0689:RMBMUA>2.0.CO;2
(a) A GOES IR 3.9-μm image of the Southeast. Urban heat islands for Nashville, Montgomery, and Atlanta are indicated as dark warm regions in the circles. (b) A contour plot of the 15-month, warm-season mean rainfall rates at a height of 2.0 km using 0.5°-resolution TRMM PR data. Values in red are greater than or equal to 4.2 mm h−1. Values in blue are less than or equal to 3.6 mm h−1
Citation: Journal of Applied Meteorology 41, 7; 10.1175/1520-0450(2002)041<0689:RMBMUA>2.0.CO;2
(a) The Georgia AEMN network of rain gauge stations. The red box represents the inset in (b). (b) An analysis of the mean rainfall amount from May to Sep (1998–2000). The redder colors are values greater than 3.0 in. (maximum value: 3.71), and bluer colors are values less than 3.0 in. (minimum value: 2.41). The values in the black oval represent possible rainfall anomalies identified by TRMM in Fig. 6b (although these values are mean total amounts, not rainfall rate)
Citation: Journal of Applied Meteorology 41, 7; 10.1175/1520-0450(2002)041<0689:RMBMUA>2.0.CO;2
(a) The Georgia AEMN network of rain gauge stations. The red box represents the inset in (b). (b) An analysis of the mean rainfall amount from May to Sep (1998–2000). The redder colors are values greater than 3.0 in. (maximum value: 3.71), and bluer colors are values less than 3.0 in. (minimum value: 2.41). The values in the black oval represent possible rainfall anomalies identified by TRMM in Fig. 6b (although these values are mean total amounts, not rainfall rate)
Citation: Journal of Applied Meteorology 41, 7; 10.1175/1520-0450(2002)041<0689:RMBMUA>2.0.CO;2
(a) The Georgia AEMN network of rain gauge stations. The red box represents the inset in (b). (b) An analysis of the mean rainfall amount from May to Sep (1998–2000). The redder colors are values greater than 3.0 in. (maximum value: 3.71), and bluer colors are values less than 3.0 in. (minimum value: 2.41). The values in the black oval represent possible rainfall anomalies identified by TRMM in Fig. 6b (although these values are mean total amounts, not rainfall rate)
Citation: Journal of Applied Meteorology 41, 7; 10.1175/1520-0450(2002)041<0689:RMBMUA>2.0.CO;2
(a) A GOES IR 3.9-μm image of Texas. Urban heat islands for Dallas, Waco, Austin, San Antonio, and Houston are indicated as warm, dark regions in the circles. (b) A contour plot of the 15-month, warm-season analysis of mean rainfall rates at a height of 2.0 km using the 0.5°-resolution TRMM PR data. Values in red are greater than or equal to 4.2 mm h−1. Values in blue are less than or equal to 3.6 mm h−1
Citation: Journal of Applied Meteorology 41, 7; 10.1175/1520-0450(2002)041<0689:RMBMUA>2.0.CO;2
(a) A GOES IR 3.9-μm image of Texas. Urban heat islands for Dallas, Waco, Austin, San Antonio, and Houston are indicated as warm, dark regions in the circles. (b) A contour plot of the 15-month, warm-season analysis of mean rainfall rates at a height of 2.0 km using the 0.5°-resolution TRMM PR data. Values in red are greater than or equal to 4.2 mm h−1. Values in blue are less than or equal to 3.6 mm h−1
Citation: Journal of Applied Meteorology 41, 7; 10.1175/1520-0450(2002)041<0689:RMBMUA>2.0.CO;2
(a) A GOES IR 3.9-μm image of Texas. Urban heat islands for Dallas, Waco, Austin, San Antonio, and Houston are indicated as warm, dark regions in the circles. (b) A contour plot of the 15-month, warm-season analysis of mean rainfall rates at a height of 2.0 km using the 0.5°-resolution TRMM PR data. Values in red are greater than or equal to 4.2 mm h−1. Values in blue are less than or equal to 3.6 mm h−1
Citation: Journal of Applied Meteorology 41, 7; 10.1175/1520-0450(2002)041<0689:RMBMUA>2.0.CO;2
URRs for the control coordinate system for five cities in the study. Blue circles are URR values in the downwind MIA. Red squares are URR values in the UCA. Green plus marks are URR values in minimum impact areas. Black asterisks are URR values over the urban area (see Fig. 1)
Citation: Journal of Applied Meteorology 41, 7; 10.1175/1520-0450(2002)041<0689:RMBMUA>2.0.CO;2
URRs for the control coordinate system for five cities in the study. Blue circles are URR values in the downwind MIA. Red squares are URR values in the UCA. Green plus marks are URR values in minimum impact areas. Black asterisks are URR values over the urban area (see Fig. 1)
Citation: Journal of Applied Meteorology 41, 7; 10.1175/1520-0450(2002)041<0689:RMBMUA>2.0.CO;2
URRs for the control coordinate system for five cities in the study. Blue circles are URR values in the downwind MIA. Red squares are URR values in the UCA. Green plus marks are URR values in minimum impact areas. Black asterisks are URR values over the urban area (see Fig. 1)
Citation: Journal of Applied Meteorology 41, 7; 10.1175/1520-0450(2002)041<0689:RMBMUA>2.0.CO;2
Summary of the downwind locations for the cities experiencing the most significant urban-impacted rainfall (shaded cross region) in the warm-season months. This analysis is based on the 3-yr, warm-season rainfall rate analysis provided by TRMM precipitation radar. The arrows represent the mean prevailing wind direction at 700 hPa as determined by the NCEP–NCAR reanalysis climatic dataset (Kalnay et al. 1996)
Citation: Journal of Applied Meteorology 41, 7; 10.1175/1520-0450(2002)041<0689:RMBMUA>2.0.CO;2
Summary of the downwind locations for the cities experiencing the most significant urban-impacted rainfall (shaded cross region) in the warm-season months. This analysis is based on the 3-yr, warm-season rainfall rate analysis provided by TRMM precipitation radar. The arrows represent the mean prevailing wind direction at 700 hPa as determined by the NCEP–NCAR reanalysis climatic dataset (Kalnay et al. 1996)
Citation: Journal of Applied Meteorology 41, 7; 10.1175/1520-0450(2002)041<0689:RMBMUA>2.0.CO;2
Summary of the downwind locations for the cities experiencing the most significant urban-impacted rainfall (shaded cross region) in the warm-season months. This analysis is based on the 3-yr, warm-season rainfall rate analysis provided by TRMM precipitation radar. The arrows represent the mean prevailing wind direction at 700 hPa as determined by the NCEP–NCAR reanalysis climatic dataset (Kalnay et al. 1996)
Citation: Journal of Applied Meteorology 41, 7; 10.1175/1520-0450(2002)041<0689:RMBMUA>2.0.CO;2
Mean 700-hPa wind direction (May–Sep) based on NCEP–NCAR reanalysis geopotential height climatic dataset from 1979 to 1998 (Kalnay et al. 1996). Mean altitude above sea level (m) of city is also given
Mean rain rates (mm h−1) from TRMM precipitation radar data (2.0-km height). The data are averaged over specified upwind, downwind, and urban area for the warm season (May–Sep) for 1998–2000. The percentage change from the upwind control area is given for the maximum impact and urban center areas. Mean percentage change for each area is also shown for all cities in the study. Mean percentage change in maximum impact area = 28.4. Mean percentage change in urban center = 5.8. Mean percentage change in northern minimum impact area = 1.1. Mean percentage change in southern minimum impact area = 10.7.
Maximum 3-yr warm-season rain rate (mm h−1) found in the maximum impact area. The table provides information on the distance from the urban center to the value in column 1. Maximum rain-rate value is found in the maximum impact area at a mean distance of ∼39 km from the edge of the urban center (or ∼64 km from the exact center)
