Abstract

The long-term trends in extreme summer season temperatures across the Twin Cities Metropolitan Area (TCMA) associated with urbanization are examined. To assess trends in extreme temperature data, maximum and minimum temperatures from 1975 to 2002 were assembled for seven stations located in both rural and urban areas. Furthermore, urbanization since 1975 was assessed by estimating the percentage of impervious surfaces from Landsat images acquired for various years. The results of this study indicated a greater rate of increase in overall minimum temperatures, resulting in a slightly declining trend in diurnal temperature range for all of the stations. In the case of extreme temperatures, most of the peripheral urban and rural stations experienced negative trends in extreme maximum temperatures, accompanied by positive trends in extreme minimum temperatures. This was also validated by the simultaneous increase in the percentage of impervious surfaces in those locations. The greatest changes were observed for Stillwater, which is located relatively close to the heart of the TCMA but has undergone a faster rate of urbanization.

1. Introduction

In recent decades there has been a rising awareness of climate change, largely due to a better understanding of long-term changes in climatic processes at different spatial scales. A leading manifestation of climate change is rising temperatures, which have widespread impacts on different components of the earth–atmosphere system. According to the fourth assessment report of the Intergovernmental Panel on Climate Change (IPCC), there has been a linear increasing trend of 0.74°C in temperatures, with the warming over the past 50 yr being nearly 2 times as much as compared with the past 100 yr (Alley et al. 2007). There is also evidence that the 11-yr period from 1995 to 2006 was the warmest on record (Alley et al. 2007). This was accompanied by regional-level warming patterns, such as the rise in the lowest daily minimum temperatures over western and central North America (Robeson 2004). In a recent study, Vose et al. (2005) analyzed global trends in maximum and minimum temperatures and diurnal temperature range (DTR) from 1979 to 2004. They found different regional-level variations in the trends of both maximum and minimum temperatures, resulting in varying trends in DTR. For instance, in the case of western North America, the DTR increased because of different rates of positive trends in maximum and minimum temperatures. However, slight declining trends in DTR were observed more in the interior of the continent over Minnesota. This increasing trend in temperatures in some of the regions has produced a greater rate of increase frequency and/or intensity in the occurrence of extreme temperatures (Easterling et al. 2000). This has been further validated by Meehl et al. (2000), who found a greater possibility in the occurrence of extreme heat stress–related events in a warmer climate.

One of the processes that influence local temperatures is associated with the degree of urbanization. This is particularly evident in the case of the United States, where half of the observed decreases in DTR were attributed to urbanization and other land use changes (Kalnay and Cai 2003). The fourth assessment report of the IPCC provided evidence of the connection between urbanization-related human activities and the increased incidence of hot days, hot nights, and heat waves in the recent years (Trenberth et al. 2007). Furthermore, over the past few decades there have been increasing concerns about the urban heat island (UHI) effect within large metropolitan regions (Arnfield and Grimmond 1998; Brazel et al. 2000). Some of the specific study sites include Malmö, Sweden (Bärring et al. 1985), Mexico City (Oke et al. 1999), Vancouver (Voogt and Grimmond 2000), Buenos Aires (Bejarán and Camilloni 2003), and Phoenix (Brazel et al. 2007).

The most predominant characteristic of a UHI is a warmer surface temperature regime across an urban area compared to surrounding rural areas. In general, UHI has been found to be most pronounced in large metropolitan areas located in the midlatitudes (Oke 1987). One major impact of urbanization is variation in microscale energy budgets in urban areas that vary at the microscale, which further shapes the spatial configuration of UHI development across urban areas (Ching et al. 1983). The main findings of most UHI-related studies show a stronger presence of the UHI effect at night than during the day, with an inverse relationship between the UHI intensity and wind speed (Souch and Grimmond 2006). In addition, there is a close relationship between the UHI development characteristics and prevailing land use, vegetation patterns, and the distribution of anthropogenic activities (Giridharan et al. 2004; Jonsson et al. 2004). Moreover, some of the studies focusing on the urban atmosphere noted a stronger attenuation in the ultraviolet portion of the electromagnetic spectrum, as in the case of Athens, Greece (Jacovides et al. 1998; Repapis et al. 1998). It has also been noted that the aerodynamic roughness of urban areas is essential for understanding UHI-related processes occurring with the urban boundary layer (Arnfield 2003). For instance, the impacts of buildings at the local level are determined by periods of exposure to solar radiation and net longwave radiation exchange (Paterson and Apelt 1989; Verseghy and Munro 1989; Arnfield 2000). At the ground level, several studies have examined the role of varied properties of different surfaces in relation to temperature, moisture, and radiation (Oke 1989; Grimmond et al. 1996).

In this context, the Twin Cities Metropolitan Area (TCMA) characterizes a typical large, midlatitude, steadily growing urban area, located in southern Minnesota near the heart of the North American continent. According to a report by the Upper Midwest Aerospace Consortium (Upper Midwest Aerospace Consortium 2008), Minnesota exhibited a warming trend between 1.5° and 3°C over the past century, which is higher than the global average. The presence of a pronounced UHI in the form of higher temperatures was first highlighted by Winkler et al. (1981), who reported the spatial concentration of higher temperatures near the most built-up areas. Their study also indicated significant differences in temperature between rural and urban stations. This was followed by another study by Ruschy et al. (1991), who reported abrupt changes in the diurnal temperature range from low winter values to high non-winter values in the case of the Twin Cities in comparison with the adjacent rural stations. These patterns can be attributed to the effects of urbanization, which cause a much greater increase in minimum temperatures than in maximum temperatures. Furthermore, Todhunter (1996) found a 2.1°C mean annual air temperature increase associated with UHI in TCMA for the year 1989, with a clear zoning in the spatial patterns of temperatures declining toward the periphery from the metropolitan core. Given the distinct effect of urbanization on temperatures, this study will update previous research by analyzing long-term trends in overall maximum and minimum temperatures, DTR, and extreme temperatures for seven stations located across the TCMA from 1975 to 2002 (Fig. 1).

Fig. 1.

Distribution of weather stations across TCMA (background map is the 2002 land cover).

Fig. 1.

Distribution of weather stations across TCMA (background map is the 2002 land cover).

With the advent of remote sensing methodology, it has become possible to monitor local urban climate changes associated with microlevel land use changes over rapidly expanding urban areas like the TCMA. Specifically, increases in urban land uses in large metropolitan areas are characterized by the spread of nonevaporative impervious surfaces, representing those materials that do not absorb water or moisture, including such urban infrastructures as rooftops, streets, highways, parking lots, and sidewalks, etc. The quantity of impervious surfaces is related to urban growth and urban density (Stankowski 1972; Rashed et al. 2005). The spatial structure of urban thermal patterns and urban heat balances are associated with urban surface characteristics. Therefore, urban imperviousness has been utilized to assess adverse influences of urbanization on urban climate, air and water quality, and natural habitat (Schueler 1994; Talmage et al. 1999; Dougherty et al. 2004). In addition, changes of urban impervious surface areas also led to the modification of energy budgets at the microclimatic scale, associated with a resultant increase in sensible heat flux at the expense of latent heat flux (Oke 1982; Owen et al. 1998). The proportion of impervious surfaces has been reported to be a good indicator for the monitoring of the spatial extent of surface UHI in urban areas. Previous studies have shown that the higher the urban size or density, the more distinctive the urban heat island effect (Oke 1976; Weng 2001). In the case of the TCMA, Yuan et al. (2005) noted a considerable increase in urban land use (from 23.7% to 32.8%) during the time period of 1986–2002 using satellite remote sensing. A positive correlation between the proportion of impervious surfaces and land surface temperatures was further indentified by Yuan and Bauer (2007) with remote sensing imagery. Moreover, the impact of change in land usage has also been linked to the near-surface atmospheric conditions. For instance, the expansion of the built-up area was found to be the main factor in long-term changes in air temperature (Huang et al. 2008). Therefore, it can be concluded that impervious surfaces in the form of urban infrastructure are associated with higher temperatures both at the surface as well as in the case of boundary layer air temperatures. Therefore, in this study we will also analyze changes in impervious surfaces around the selected network of stations and their association with trends in extreme temperatures and DTR for the same time period.

2. Study area

The TCMA consists of the seven counties of Anoka, Carver, Dakota, Hennepin, Ramsey, Scott, and Washington. They contain a diversity of human activities with an agglomeration of dense urban land uses mostly located in the metropolitan core, and rural land uses still prevailing in the peripheral areas of TCMA. Topographically, the TCMA offers an advantageous site for urban climate studies with minimal local relief (Todhunter 1996). The TCMA has steadily expanded outward since the 1960s, and its population has increased by 38% during 1974–2000 (EPA 1997), with a simultaneous increase of 59% in urban land area over the same time period (Yuan and Bauer 2007). Mean imperviousness has increased from approximately 9% in 1986 to 14% in 2000 (Bauer et al. 2004). Increased percentage of impervious surface is associated with increased mean land surface temperature (Yuan and Bauer 2007) as well as average water temperature in the TCMA (Talmage et al. 1999). As noted above, the presence of a pronounced UHI has also led to the occurrence of considerably higher temperatures in the most urbanized parts of the metropolitan area (Winkler et al. 1981; Todhunter 1996). In the present study, we examine the impact of change in impervious surface on trends in near-surface air temperatures as well as long-term trends in extreme temperatures and DTR over the past 3 decades.

3. Datasets

Daily station-level climate data for maximum and minimum temperatures were obtained from the widely used Global Daily Climatology Network (GDCN) dataset 9101 v 1.0. This GDCN dataset was consolidated from data collected over several years of analysis as well as the archiving of daily data into a single dataset by the National Climatic Data Center (NCDC). All daily climatic data from GDCN have been checked through an extensive set of quality control measures including datum checks and statistical analyses of sets of observations to locate and identify possible outliers and erroneous data. Additional details concerning the methodology and quality control techniques used to maintain the accuracy of the GDCN dataset are described in Gleason (2002).

There are a total of 14 stations covering the entire study area. However, in the final analysis, only seven stations were used, mainly in view of the presence of continuous records for most of the study period extending from 1975 to 2002 during the summer months of June, July, and August (Fig. 1). On average, missing data totaled less than 5% for the entire study period for both maximum and minimum temperatures. Among the selected stations, five stations (Cedar, Farmington, Jordan, Rosemount, and Stillwater) are located in the suburban ring, while the remaining two stations (Minneapolis and St. Paul) are located in the core cities. The mixes of both urban and rural locations are important for determining the influence of urbanization on long-term trends in temperature patterns.

Urbanization was assessed by measuring changes of percent of impervious surface areas in the TCMA. To calculate the percent imperviousness change over time, five multitemporal Landsat scenes were acquired for 5 yr during a 30-yr time span: 1975, 1986, 1991, 1998, and 2002. Except for the 1975 data, which are a multispectral scanner (MSS) image with 80-m resolution, all the other images are Landsat Thematic Mapper (TM) with 30-m resolution for multispectral bands. All images were rectified and projected to UTM zone 15, GRS1980, NAD83, with geometric errors amounting to less than one-fourth of pixel size. The original digital numbers of the image were also converted to exoatmospheric reflectance.

4. Analysis and results

The results of these analyses are described in the following three sections.

a. Trends of average summer maximum and minimum temperatures

The average maximum temperatures during the study period ranged from 23° to 33°C, with considerably little variation among the seven stations during the study period (Fig. 2a). The highest maximum temperature was observed in Rosemount (32.2°C in 1988), and the lowest temperature was observed in Jordan (23.3°C in 1992). Conversely, in the case of average summer minimum temperatures, there were considerable year-to-year variations during the study period (Fig. 2b). The maximum variations were observed in the cases of Rosemount, Cedar, and Jordan, with relatively less variability observed in the case of the two central-city stations in St. Paul and Minneapolis. The minimum temperatures ranged from 9° to 18°C over the study period. A detailed examination of the minimum temperatures reveals the occurrence of substantially lower temperatures in Jordan during 1987–89. This can be explained from the examination of the station history provided by the National Weather Service, which revealed that the station in Jordan was moved from 932 to 930 ft in 1987. It was again moved to a slightly different location in 1990. Furthermore, the instruments used to measure temperatures were changed from maximum and minimum thermometer to the Maximum–Minimum Thermometer System (MMTS) electronic sensor.

Fig. 2.

Plot of average summer temperatures during 1975–2002 for seven stations spread across the study area: (a) maximum temperatures and (b) minimum temperatures.

Fig. 2.

Plot of average summer temperatures during 1975–2002 for seven stations spread across the study area: (a) maximum temperatures and (b) minimum temperatures.

The overall trends in maximum temperatures were in general negative, except in the case of Stillwater where the trend was +0.2°C per decade (Fig. 3). The largest overall decline in maximum temperatures was observed in case of Cedar at −1.1°C per decade. It is noteworthy that most of the trends in the case of maximum temperatures were not statistically significant except in the case of Cedar (Table 1). However, in the case of minimum temperatures the trends were reversed with all the stations experiencing positive trends (Fig. 3). The smallest increase was observed in the Twin Cities themselves, with the greatest increase of +0.8°C per decade observed for Stillwater, followed by Rosemount (+0.73°C per decade), both of which have undergone rapid urbanization during the past decade. The trends in the case of minimum temperatures were significant at a 0.05 or higher confidence level for almost all stations except for the two major urban stations of St. Paul and Minneapolis, as well as Cedar (Table 1).

Fig. 3.

Overall trends calculated as regression coefficients in maximum, minimum, and mean temperatures and DTR from 1975 to 2002.

Fig. 3.

Overall trends calculated as regression coefficients in maximum, minimum, and mean temperatures and DTR from 1975 to 2002.

Table 1.

Significance-level values for calculated regression coefficients for maximum and minimum temperatures, and DTR. Boldface values indicate significance at ≥0.05 confidence level.

Significance-level values for calculated regression coefficients for maximum and minimum temperatures, and DTR. Boldface values indicate significance at ≥0.05 confidence level.
Significance-level values for calculated regression coefficients for maximum and minimum temperatures, and DTR. Boldface values indicate significance at ≥0.05 confidence level.

We further examined the trend in DTR across the seven stations. The trends were evidently negative for all stations, with the lowest negative trend observed for Minneapolis at −0.48°C per decade. The maximum decline was observed in Cedar (−1.5°C per decade). In the case of DTR, the trends were significant at a 0.05 confidence level for all stations, with the highest statistical significance observed in the cases of Cedar, Farmington, and Rosemount (Table 1).

b. Trends in temperature patterns in the extreme intensity categories

In view of the increasing awareness of global warming, a growing number of studies in recent years have focused on extreme temperatures (Bell et al. 2004; Vincent and Mekis 2006). Several of these studies have linked the impact of UHI-related processes to extreme temperatures (Balling et al. 1990; Yan et al. 2002). To detect long-term trends in temperature patterns in the extreme intensity categories, threshold values were calculated for the extreme two percentile categories that correspond to the 90th and 95th percentile levels at each station.

The threshold values for each percentile category were calculated by sorting the data for all years in ascending order and then calculating the threshold value for the top 5 and 10 percentile categories separately for each station. Once the threshold values for a station were determined, each year was checked separately to enumerate the number of days falling within each percentile category. The values in each of the percentile categories were exclusive to that specific percentile interval and not a cumulative score for all categories below it. Two variables were examined to reveal trends in extreme temperature events, which include frequency of events and the average of the total number of events occurring above the threshold for each year.

Specifically, frequency consists of the number of days for each year and for each station, when the observed temperatures were higher than the identified threshold value. On the other hand, the average value consists of the mean value of the observed temperatures for all the days in each extreme percentile category. This resulted in individual station-level matrices, consisting of 28 rows (1 row for each year), and 2 columns (1 column each for frequency and average temperatures within each extreme percentile category). Next, Pearson product moment correlation coefficients were calculated using the year as the independent variable and the maximum or minimum temperatures as the dependent variable for the 90th and 95th percentile categories. The trends in the two extreme percentile categories were calculated separately for frequency and average temperatures in the case of both maximum and minimum temperatures. Finally, there were two matrices consisting of seven rows representing the weather stations, with four columns consisting of the trend values for the extreme percentile intervals for the average and frequency of temperature incidences.

Figure 4 shows the trends in the occurrence of frequency and average summer maximum temperatures for the upper two extreme five-percentile interval categories. In general, except in the case of Stillwater, negative trends of frequency of extreme temperature were found, and the trends were substantially higher in the case of average maximum temperatures (Fig. 4b) than those observed in the case of frequency of temperatures (Fig. 4a) for most stations. The two central-city stations—Minneapolis and St. Paul—experienced negative trends in frequency of events for both the extreme percentile categories. The trends were more prominent for the 90th percentile in comparison to the 95th percentile (Fig. 4a). The three rural stations—Cedar, Jordan, and Rosemount—experienced higher rates of decline in both frequency and average temperatures in both percentile categories in comparison with other stations. Jordan experienced the highest decline in the 90th percentile category at −4°C per decade (Fig. 4b). Cedar located to the north of the study area experienced the largest decrease in the case of the 95th percentile at −4.6°C per decade. The average temperatures in the 95th percentile category showed a slightly positive trend of +0.4°C per decade in the case of St. Paul. Stillwater, particularly, which is located to the east of the study area, experienced positive trends in both frequency and average of extreme maximum temperatures. The statistical significance values associated with the regression coefficients are shown in Table 2. The results of the above analysis in the case of maximum temperatures were significant at less than the 0.05 level of confidence for both the extreme percentile categories for frequency of maximum temperatures, in the case of Cedar, and for frequency of events in the 90th percentile category in Jordan, Farmington, and St. Paul.

Fig. 4.

Trends (1975–2002) expressed as the Pearson product moment correlation also expressed as the regression coefficients for maximum temperatures occurring in the extreme upper two five-percentile class intervals: (a) frequency of days and (b) average temperatures. Values plotted for each percentile interval are exclusive to that category.

Fig. 4.

Trends (1975–2002) expressed as the Pearson product moment correlation also expressed as the regression coefficients for maximum temperatures occurring in the extreme upper two five-percentile class intervals: (a) frequency of days and (b) average temperatures. Values plotted for each percentile interval are exclusive to that category.

Table 2.

Significance-level values for calculated regression coefficients for the two extreme percentile categories for maximum temperatures. Boldface values indicate significance at ≥0.05 confidence level.

Significance-level values for calculated regression coefficients for the two extreme percentile categories for maximum temperatures. Boldface values indicate significance at ≥0.05 confidence level.
Significance-level values for calculated regression coefficients for the two extreme percentile categories for maximum temperatures. Boldface values indicate significance at ≥0.05 confidence level.

For minimum temperatures, however, most of the stations experienced positive trends in both extreme percentile categories (Fig. 5). In the case of the rural stations there was a slight downward trend for average temperatures in the 95th percentile interval category. Jordan and Rosemount located in the southern half of the study area experienced relatively higher positive trends, particularly in the case of the frequency of events. Farmington and Rosemount experienced considerably higher positive trends for both frequency and average temperature events in both the extreme percentile categories. Whereas the trend in minimum temperatures was generally positive for most of the stations, the twin cities experienced negative trends for both frequency and average temperature events for the extreme percentile categories, except for frequency of events in the 90th percentile category in Minneapolis. The results were statistically significant above the 0.10 level in the case of Stillwater for both frequency and average minimum temperatures observed in the 95th percentile level, and for frequency of extreme temperatures in the 90th percentile category in the cases of Jordan and Rosemount (Table 3).

Fig. 5.

Trends in minimum temperatures for the extreme upper two five-percentile categories: (a) frequency of days and (b) average temperatures, expressed as in Fig. 4.

Fig. 5.

Trends in minimum temperatures for the extreme upper two five-percentile categories: (a) frequency of days and (b) average temperatures, expressed as in Fig. 4.

Table 3.

Significance-level values for calculated regression coefficients for the two extreme percentile categories for minimum temperatures. Boldface values indicate significance at >0.10 confidence level.

Significance-level values for calculated regression coefficients for the two extreme percentile categories for minimum temperatures. Boldface values indicate significance at >0.10 confidence level.
Significance-level values for calculated regression coefficients for the two extreme percentile categories for minimum temperatures. Boldface values indicate significance at >0.10 confidence level.

The declining trends in extreme maximum temperatures accompanied by increases in extreme minimum temperatures can be attributed to the modulating role of urban structures as heat absorbing structures. Positive trends in minimum temperatures for most of the stations may be attributed to the role of heat released from impervious surfaces, such as paved surface and concrete buildings. It can also be attributed to the increasing population over the study period, resulting in a more significant anthropogenic influence on long-term trends in temperatures. It has been noted in the latest IPCC report that places experiencing rapid urbanization will be more prone to extreme weather events with clear evidence of the role of anthropogenic activities on long-term climate change (Trenberth et al. 2007). The overall decline in extreme maximum and minimum temperatures for the two core stations located in the Twin Cities conforms to other studies examining extreme temperature days in large metropolitan areas like Beijing (Koo 1988; Koo and Chang 1989).

c. Trends of extreme temperatures in relation to changes of percentage imperviousness

Given clear evidence of an increasing trend in extreme temperatures in some of the less urbanized and relatively rural stations in the study area, we examined the change in the percentage of impervious surface around the station sites during the study period. The normalized spectral mixture analysis (NSMA) was used to calculate the percent of impervious surface areas. The NSMA adds a normalization process before applying linear spectral mixture analysis (LSMA), which models a mixed spectrum as a linear combination of three end members, vegetation-impervious surface soil. The purpose of normalization is to reduce within-class radiometric variations while maintaining necessary information to separate major end members (Wu 2004). More detailed information about the NSMA method can be found in Wu (2004) and Wu and Yuan (2007).

Next, the mean percentage imperviousness maps were generated by using NSMA for two buffer zones of 0.5 and 8 km, respectively, surrounding each of the seven stations for five different years. The changes in imperviousness around each station are displayed in Fig. 6. These buffer zones are representative of physical processes occurring at different spatial scales within the urban canopy layer (UCL). Specifically, the 0.5-km buffer zone is representative of the block level comprising blocks and factories at the microscale (Oke 2006). On the other hand, the 8-km buffer zone is approximately representative of the urban terrain zone (Ellefsen 1991) or urban climate zone (Oke 2004) at the meso- or local scale. Therefore, the 0.5-km buffer (Fig. 6a) is more useful for understanding urbanization-related dynamics in the immediate vicinity of the selected seven stations. The 8-km buffer (Fig. 6b) is more useful for considering the description of urban structure, which is relevant for roughness, airflow, radiation access, and screening (Oke 2006). The clear rising tendency in percent imperviousness for all the stations in Fig. 6b indicates continuous and rapid urban growth over the 28-yr study period. The slopes of the curves in Fig. 6a show dramatic developments for Farmington and Minneapolis after 1991. By visually checking the satellite images, we found that newly developed urban land use surrounding Farmington is largely residential. On the other hand, the increased imperviousness value for the Minneapolis station is due to the expansion of the Minneapolis–St. Paul International Airport. The rapid urban growth in Farmington may be associated with the steeper increase of minimum temperatures at this station because greater amounts of heat are being stored by impervious surfaces within the urban structures.

Fig. 6.

Average percent imperviousness of the (a) 0.5- and (b) 8-km buffer areas surrounding the seven stations from 1975 to 2002.

Fig. 6.

Average percent imperviousness of the (a) 0.5- and (b) 8-km buffer areas surrounding the seven stations from 1975 to 2002.

Furthermore, given the sharp increase in the percentage of impervious surface since 1991 (see Fig. 6b), we also calculated the trends separately between 1991 and 2002 and found these trends are positive and higher than the overall trends for all the urban stations in the case of extreme minimum temperatures. This may be attributed to the fact that impervious surfaces can affect long-term temperature trends at the mesoscale by storing more heat and waterproofing of the surface, thereby modifying the energy budget. In addition, the blocklike geometry of urban structures creates greater possibilities for trapping heat as well as stagnation of air within the canopy layer (Oke 1987). For instance, the increase in imperviousness in the TCMA involves the increase in surface roughness in the form of high-rise buildings. As a result, there is a considerable decline in albedo with radiation getting trapped inside the UCL, which leads to higher levels of counterradiation.

5. Discussion and conclusions

Our analysis focused on overall trends in temperatures as well as trends in extreme intensity categories in response to urbanization in the TCMA. The extreme intensity categories were defined by the frequency and average temperatures in the top two five-percentile categories. The main findings of our analysis are below.

  • In the case of the overall trends in maximum temperatures, all the stations except Stillwater experienced negative trends. The smallest decreases were observed for the case of Farmington, a rapidly growing suburban city, and the two core cities of TCMA.

  • The overall trends for minimum temperatures were positive for all stations. The smallest increases were again observed in the central cities, while all the rural stations experienced a greater rate of increase in minimum temperatures.

  • Most of the rural stations experienced negative trends in extreme maximum temperatures and frequencies, while positive trends were observed for extreme minimum temperatures.

  • In the case of the two central-city stations, trends in extreme temperatures and frequencies were mostly negative.

  • The clear increasing percentage imperviousness for all seven stations indicates a continuous and persistent urbanization process over the 28-yr period. Furthermore, the increase in percentage impervious surface areas is greater in the rural and suburban areas than that observed for the urban core areas.

  • The role of impervious surfaces on temperature is clearly visible in the form of higher rates of increase in extreme temperatures in the three stations of Stillwater, Farmington, and Rosemount, which experienced the greatest increases in percentage of impervious surface areas.

In summary, it is evident that the steady expansion of urban land uses has impacted long-term trends in microclimatic conditions across TCMA. The spatial patterns in the trends of extreme temperatures were more prominent over the suburban stations located near the metropolitan core that have undergone greater rates of land use change through the recent past. Trends were less clear for the core cities of Minneapolis and St. Paul. Furthermore, in the cases of Stillwater, Rosemount, and Farmington, the trends were stronger during the past 10 yr, which corresponds to the period of greater increase in percentage of impervious surfaces. The expansion of impervious surfaces in TCMA has led to a greater heat storing capacity, thereby leading to higher levels of counterradiation around the urban stations. This is specifically evident for transforming suburban stations. It can be concluded from our analysis that the role of UHI can be seen more clearly for stations in the immediate vicinity of the two urban core stations, with the influence declining with distance. This is particularly evident from the steep increase of percentage of impervious surface within the 8-km buffer zone around Stillwater, Farmington, and Rosemount, which also showed the greatest increases in extreme minimum temperatures over the study period. Furthermore, the results of this study will be particularly helpful for planners in developing scenarios for future land cover changes across the TCMA.

Acknowledgments

We express our gratitude to Dr. P. O. Muller for his careful editing of the final manuscript, which made the paper much easier to read. We are also grateful for the useful suggestions provided by the anonymous reviewers and the editor.

REFERENCES

REFERENCES
Alley
,
R.
, and
Coauthors
,
2007
:
Climate Change 2007: The Physical Science Basis: Summary for Policymakers.
IPCC, 21 pp
.
Arnfield
,
A. J.
,
2000
:
A simple model of urban canyon energy budget and its validation.
Phys. Geogr.
,
21
,
305
326
.
Arnfield
,
A. J.
,
2003
:
Two decades of urban climate research: A review of turbulence, exchanges of energy and water, and the urban heat island.
Int. J. Climatol.
,
23
,
1
26
.
Arnfield
,
A. J.
, and
C. S. B.
Grimmond
,
1998
:
An urban canyon energy budget model and its application to urban storage heat flux modeling.
Energy Build.
,
27
,
61
68
.
Balling
,
R. C.
,
J. A.
Skindlov
, and
D. H.
Phillips
,
1990
:
The impact of increasing summer mean temperatures on extreme maximum and minimum temperatures in Phoenix, Arizona.
J. Climate
,
3
,
1491
1494
.
Bärring
,
L.
,
J. O.
Mattsson
, and
S.
Lindqvist
,
1985
:
Canyon geometry, street temperatures and urban heat island in Malmö, Sweden.
Int. J. Climatol.
,
5
,
433
444
.
Bauer
,
M. E.
,
N. J.
Heinert
,
J. K.
Doyle
, and
F.
Yuan
,
2004
:
Impervious surface mapping and change monitoring using Landsat remote sensing.
Proc. American Society of Photogrammetry and Remote Sensing Annual Conf., Denver, CO, American Society of Photogrammetry and Remote Sensing, 10 pp. [Available online at http://land.umn.edu/documents/ASPRS%202004_Bauer_Impervious%20mapping.pdf.]
.
Bejarán
,
R. A.
, and
I. A.
Camilloni
,
2003
:
Objective method for classifying air masses: An application to the analysis of Buenos Aires’ (Argentina) urban heat island intensity.
Theor. Appl. Climatol.
,
74
,
93
103
.
Bell
,
J. L.
,
L. C.
Sloan
, and
M. A.
Snyder
,
2004
:
Regional changes in extreme climatic events: A future climate scenario.
J. Climate
,
17
,
81
87
.
Brazel
,
A. J.
,
N.
Selover
,
R.
Vose
, and
G.
Heisler
,
2000
:
The tale of two climates—Baltimore and Phoenix urban LTER sites.
Climate Res.
,
15
,
123
135
.
Brazel
,
A. J.
,
P.
Gober
,
S-J.
Lee
,
S.
Grossman-Clarke
,
J.
Zehnder
,
B.
Hedquist
, and
E.
Comparri
,
2007
:
Determinants of changes in the regional urban heat island in metropolitan Phoenix (Arizona, USA) between 1990 and 2004.
Climate Res.
,
33
,
171
182
.
Ching
,
J. K. S.
,
J. F.
Clarke
, and
J. M.
Godowitch
,
1983
:
Modulation of heat flux by different scales of advection in an urban environment.
Bound.-Layer Meteor.
,
25
,
171
191
.
Dougherty
,
M.
,
R. L.
Dymond
,
S. J.
Goetz
,
C. A.
Jantz
, and
N.
Goulet
,
2004
:
Evaluation of impervious surface estimates in a rapidly urbanizing watershed.
Photogramm. Eng. Remote Sens.
,
70
,
1275
1284
.
Easterling
,
D. R.
,
J. L.
Evans
,
P. Ya
Groisman
,
T. R.
Karl
,
K. E.
Kunkel
, and
P.
Ambenje
,
2000
:
Observed variability and trends in extreme climate events: A brief review.
Bull. Amer. Meteor. Soc.
,
81
,
417
425
.
Ellefsen
,
R.
,
1991
:
Mapping and measuring buildings in the urban canopy boundary layer in ten US cities.
Energy Build.
,
15–16
,
1025
1049
.
EPA
,
1997
:
Climate change and Minnesota.
EPA 230-F-97-008w, 4 pp
.
Giridharan
,
R.
,
S.
Ganesan
, and
S. S. Y.
Lau
,
2004
:
Daytime urban heat island effect in high-rise and high-density residential developments in Hong Kong.
Energy Build.
,
36
,
525
534
.
Gleason
,
B. E.
,
2002
:
Data documentation for data set 9101: Global daily climatology network, V1.0 July 22, 2002.
National Climatic Data Center, 26 pp
.
Grimmond
,
C. S. B.
,
C.
Souch
, and
M. D.
Hubble
,
1996
:
Influence of tree cover on summertime surface energy balance fluxes, San Gabriel Valley, Los Angeles.
Climate Res.
,
6
,
45
57
.
Huang
,
L.
,
J.
Lia
,
D.
Zhaoa
, and
J.
Zhub
,
2008
:
A fieldwork study on the diurnal changes of urban microclimate in four types of ground cover and urban heat island of Nanjing, China.
Build. Environ.
,
43
,
7
17
.
Jacovides
,
C. P.
,
G. P.
Gianourakos
,
D. N.
Asimakopoulos
, and
M. D.
Steven
,
1998
:
Measured spectra of solar ultraviolet irradiances at Athens basin, Greece.
Theor. Appl. Climatol.
,
59
,
107
119
.
Jonsson
,
P.
,
C.
Bennet
,
I.
Eliasson
, and
E.
Selin Lindgren
,
2004
:
Suspended particulate matter and its relations to the urban climate in Dar es Salaam, Tanzania.
Atmos. Environ.
,
38
,
4175
4181
.
Kalnay
,
E.
, and
M.
Cai
,
2003
:
Impact of urbanization and land-use change on climate.
Nature
,
423
,
528
531
.
Koo
,
E. H.
,
1988
:
Climate change: Secular trends in urban temperature.
Hong Kong Observatory Publication Reprint 172
.
Koo
,
E. H.
, and
W. L.
Chang
,
1989
:
Statistical analysis of temperatures in Hong Kong.
Hong Kong Observatory Publication Reprint 212
.
Meehl
,
G. A.
, and
Coauthors
,
2000
:
Trends in extreme weather and climate events: Issues related to modeling extremes in projections of future climate change.
Bull. Amer. Meteor. Soc.
,
81
,
427
436
.
Oke
,
T. R.
,
1976
:
City size and the urban heat island.
Atmos. Environ.
,
7
,
769
779
.
Oke
,
T. R.
,
1982
:
The energetic basis of the urban heat island.
Quart. J. Roy. Meteor. Soc.
,
108
,
1
24
.
Oke
,
T. R.
,
1987
:
Boundary Layer Climates.
2nd ed. Routledge, 450 pp
.
Oke
,
T. R.
,
1989
:
The micrometeorology of the urban forest.
Philos. Trans. Roy. Soc.
,
324
,
335
349
.
Oke
,
T. R.
,
2004
:
Initial guidance to obtain representative meteorological observations at urban sites.
IOM Rep. 81, World Meteorological Organization, 51 pp
.
Oke
,
T. R.
,
2006
:
Towards better scientific communication in urban climate.
Theor. Appl. Climatol.
,
84
,
179
190
.
Oke
,
T. R.
,
R. A.
Spronken-Smith
,
E.
J’auregui
, and
C. S. B.
Grimmond
,
1999
:
The energy balance of central Mexico City during the dry season.
Atmos. Environ.
,
33
,
3919
3930
.
Owen
,
T. W.
,
T. N.
Carlson
, and
R. R.
Gillies
,
1998
:
An assessment of satellite remotely-sensed land cover parameters in quantitatively describing the climatic effect of urbanization.
Int. J. Remote Sens.
,
19
,
1663
1681
.
Paterson
,
D. A.
, and
C. J.
Apelt
,
1989
:
Simulation of wind flow around three-dimensional buildings.
Build. Environ.
,
24
,
39
50
.
Rashed
,
T.
,
J. R.
Weeks
,
D.
Stow
, and
D.
Fugate
,
2005
:
Measuring temporal compositions of urban morphology through spectral mixture analysis: Toward a soft approach to change analysis in crowded cities.
Int. J. Remote Sens.
,
26
,
699
718
.
Repapis
,
C. C.
,
H. T.
Mantis
,
A. G.
Paliatsos
,
C. M.
Philandras
,
A. F.
Bais
, and
C.
Meleti
,
1998
:
Case study of UV-B modification during episodes of urban air pollution.
Atmos. Environ.
,
32
,
2203
2208
.
Robeson
,
S. M.
,
2004
:
Trends in time-varying percentiles of daily minimum and maximum temperature over North America.
Geophys. Res. Lett.
,
31
,
L04203
.
doi:10.1029/2003GL019019
.
Ruschy
,
D. L.
,
D. G.
Baker
, and
R. H.
Skaggs
,
1991
:
Seasonal variation in daily temperature ranges.
J. Climate
,
4
,
1211
1216
.
Schueler
,
T. R.
,
1994
:
The importance of imperviousness.
Watershed Prot. Technol.
,
1
,
100
111
.
Souch
,
C.
, and
S.
Grimmond
,
2006
:
Applied climatology: Urban climate.
Prog. Phys. Geogr.
,
30
,
270
279
.
Stankowski
,
S. J.
,
1972
:
Population density as an indirect indicator of urban and suburban land-surface modifications.
Professional Paper, Rep. 800B, U.S. Geological Survey, 6 pp
.
Talmage
,
P. J.
,
K. E.
Lee
,
R. M.
Goldstein
,
J. P.
Anderson
, and
J. D.
Fallon
,
1999
:
Water quality, physical habitat, and fish-community composition in streams in the Twin Cities Metropolitan Area, Minnesota 1997–98.
U.S. Geological Survey, 99-4247, 18 pp
.
Todhunter
,
P. E.
,
1996
:
Environmental indices for the twin cities metropolitan area (Minnesota, USA) urban heat island— 1989.
Climate Res.
,
6
,
59
69
.
Trenberth
,
K. E.
, and
Coauthors
,
2007
:
Observations: Surface and atmospheric climate change.
Climate Change 2007: The Physical Science Basis, S. Solomon et al., Eds., Cambridge University Press, 235–336
.
Upper Midwest Aerospace Consortium
,
2008
:
Climate maps: Temperature and precipitation averages.
.
Verseghy
,
D. L.
, and
D. S.
Munro
,
1989
:
Sensitivity studies on the calculation of the radiation balance of urban surfaces. Part I: Shortwave radiation.
Bound.-Layer Meteor.
,
46
,
309
331
.
Vincent
,
L. A.
, and
E.
Mekis
,
2006
:
Changes in daily and extreme temperature and precipitation indices for Canada over the twentieth century.
Atmos.–Ocean
,
44
,
177
193
.
Voogt
,
J. A.
, and
C. S. B.
Grimmond
,
2000
:
Modeling surface sensible heat flux using surface radiative temperatures in a simple urban area.
J. Appl. Meteor.
,
39
,
1679
1699
.
Vose
,
R. S.
,
D. R.
Easterling
, and
B.
Gleason
,
2005
:
Maximum and minimum temperature trends for the globe: An update through 2004.
Geophys. Res. Lett.
,
32
,
L23822
.
doi:10.1029/2005GL024379
.
Weng
,
Q.
,
2001
:
A remote sensing–GIS evaluation of urban expansion and its impact on surface temperature in the Zhujiang Delta, China.
Int. J. Remote Sens.
,
22
,
1999
2014
.
Winkler
,
J. A.
,
R. H.
Skaggs
, and
D. G.
Baker
,
1981
:
Effect of temperature adjustments on the Minneapolis-St. Paul Urban Heat Island.
J. Appl. Meteor.
,
20
,
1295
1300
.
Wu
,
C.
,
2004
:
Normalized spectral mixture analysis for monitoring urban composition using ETM+ imagery.
Remote Sens. Environ.
,
93
,
480
492
.
Wu
,
C.
, and
F.
Yuan
,
2007
:
Seasonal sensitivity analysis of impervious surface estimation with satellite imagery.
Photogramm. Eng. Remote Sens.
,
73
,
1393
1401
.
Yan
,
Z.
, and
Coauthors
,
2002
:
Trends of extreme temperatures in Europe and China based on daily observations.
Climatic Change
,
53
,
355
392
.
Yuan
,
F.
, and
M. E.
Bauer
,
2007
:
Comparison of impervious surface area and normalized difference vegetation index as indicators of surface urban heat island effects in Landsat imagery.
Remote Sens. Environ.
,
106
,
375
386
.
Yuan
,
F.
,
K. E.
Sawaya
,
B. C.
Loeffelholz
, and
M. E.
Bauer
,
2005
:
Land cover classification and change analysis of the Twin Cities (Minnesota) Metropolitan Area by multitemporal Landsat remote sensing.
Remote Sens. Environ.
,
98
,
317
328
.

Footnotes

Corresponding author address: Shouraseni Sen Roy, Dept. of Geography and Regional Studies, University of Miami, 225 Ferre Building, 1000 Memorial Drive, Coral Gables, FL 33146. Email: ssr@miami.edu