The Urban Moisture Climate

by DOUGLAS L. SISTERSON and RICHARD A. DIRKS, graduate research assistant and professor, Department of Atmospheric Science, University of Wyoming, Laramie, Wyo. Sisterson is now with the Atmospheric Physics Section, Argonne National Labora- tory, Argonne, Ill.

ABSTRACT.-Data collected on 26 July 1974 as a part of project METROMEX in St. Louis show the three-dimensional structure of the urban moisture field. Mesoscale dry regions at the urban surface, corresponding to large residential and light industrial land-use characterization, were responsible for a reduction in specific humidity in the urban mixing layer. Anthropogenic sources and sinks of water vapor were estimated and found to be insignificant for summer daytime mesoscale influences on the atmospheric water balance of the urban area. The effects of reduced moisture in the urban area are directly related to the urban heat island, visibility, fog, precipitation, and human comfort in our cities.

HUMANLIFE PATTERNS have reduction in latent heat energy, may changed radically in the last century also account for anomalous precipitation with the increasing migration from patterns (Terjung et al. 1971; Boatman rural to urban areas. The increased 1974). On the local scale uncertainty magnitude of climatic anomalies ac- has been expressed whether the overall companying urban-industrial expansion water contribution from the city is re- are well documented by the urban heat duced compared to rural areas since island and are further evident in the effects on visibility, fog, moisture, and much water vapor is released by com- precipitation. These factors are im- bustion, industry, and cooling operations. portant to human comfort, land-use Project METROMEX (METROpoli- planning, and engineering considera- tan Meteorological Experiment) was tions. initiated in 1971 to statistically and The moisture budget appears to be a physically evaluate the possible modifi- key contributor to urban climate ano- cation of weather and climate by an malies. Seasonal surface-moisture ano- urban-industrial complex. Recent studies malies have been noted in large urban by the University of Wyoming in Proj- areas. The greatest urban-rural humid- ect METROMEX focused on the prob- ity differences are during summer after- lem of urban moisture (Sisterson 1975). noons when the urban dry island is Data were acquired from the instru- greatest in magnitude. Summer night- mented Wyoming research aircraft and time differences in urban-rural absolute instrumented meteorological weather humidity patterns suggest that the city mobiles, along with surface networks is drier when the nocturnal heat island operated by cooperating groups (Illinois is strongest. Wintertime studies, how- State Water Survey, St. Louis City and ever, show that the urban area is more County Air Pollution Agencies). moist than its rural counterpart. In- Three aspects of urban moisture were tensification of the urban heat island has investigated : (1) relative magnitudes been directly related to less absolute of anthropogenic sources and sinks, (2) moisture (Wood 1971) and reduced eva- the three-dimensional structure of the poration (Myrup 1969). A reduction in moisture field, and (3) evaluation of urban moisture, with a corresponding the moisture budget components. SOURCES AND SINKS OF MOISTURE precipitation days, so the precipitation component was neglected. Particulate Estimates of automobile and in- mass concentrations of 10OPgm m-3 dustrial sources of water vapor were (0.5pm size) have been reported for a compared to ~otentialevaporation cal- highly polluted day in Los Angeles, culations. The industries chosen in this California. In the extreme case, all study warranted investigation because particles are assumed to be hygroscopic of their size and location in the St. Louis salt particles at a relative humidity of area. The results of water vapor pro- 90 percent. For this case, the maximum duction rates (gm hr-') for anthropo- change in specific humidity in the urban genic sources and potential evaporation area (30 km by 30 km by 1.5 km) would for the urban area are shown in table 1. be only 2.75 x 10-4 gm kgm-I. This Even if only a fraction of the radiation amount is insignificant and cannot be energy is used for evaporation, it is ap- detected by moisture measurement. parent that potential evaporation (for a Investigation of photochemical re- 30-km by 30-km surface area) is several actions in the urban atmosphere was orders of magnitude greater than an- limited to the formation of sulfuric acid, thropogenic moisture sources. However, the most common and prolific photo- industrial sources are essentially point chemical pollutant. In the extreme case, sources of water vapor. If the potential all sulfur dioxide was assumed to react evaporation is considerd for an area with water vapor to form sulfuric acid. only 3 km by 3 km, the rate of water In all cases, the water vapor necessary vapor production approaches that of to convert all the SO, to HISO, was at the major industrial sources. Thus, least one order of magnitude less than during the daylight hours, anthropo- the water vapor emissions from the genic sources do not appear to be of industry. consequence except on the local scale. Sulfuric acid is also deliquescent at At nighttime, when potential evapora- high concentrations. The excess water tion is greatly reduced and mixing is vapor between industrial source and re- limited, anthropogenic sources are in- action requirements was used to dilute deed a relevant factor in the urban sulfuric acid concentrations in the in- water balance, so that the urban atmo- dustrial plumes. The Labadie power sphere at night may be more moist than plant plume showed the greatest poten- the rural atmosphere. tial concentrations of sulfuric acid. Only Hygroscopic aerosols, photochemical if the environmental relative humidity reactions, and precipitation are major is greater than 94 percent could sulfuric moisture sinks for an urban atmosphere. acid from the Labadie plant absorb In this study, cases were chosen on non- ambient water vapor--again assuming

Table I.-Urban area sources of water vapor

Classification Source Water vapor (gm hr-I) Auto Automobile Refineries Amoco Shell Power Plants Meramec Portage des Sioux Labadie Cement Alfa Missouri Portland Steel Granite City Nature Potential Evaporation -- all SO, converted to H,SO,. The photo- served on a number of fair summer chemical reaction of sulfuric dioxide and days. Low-level winds on this day were water vapor is therefore clearly not a northerly with speeds less than 5 m significant factor in the extraction of sec-l. Specific humidity and potential water vapor from the urban atmo- temperature patterns are shown for the sphere on the mesoscale, and in general, surface at 1600 CDT (fig. 1). Dry re- the deliquescense of sulfuric acid is gions exist along the northern, southern, insignificant. and western edges of St. Louis, corresponding to major residential and ST. LOUIS CASE STUDY- light industrial land-use characteriza- 26 JULY 1974 tions (fig. 2). Heat islands are apparent north of the city and along the southern Atmospheric moisture patterns are in- edge. Because these analyses are based fluenced by advection, turbulent mixing, upon individual stations, local effects and local sources and sinks of moisture. influence station readings, and small In the METROMEX field studies, air- regions of low moisture and high temp- craft transects were chosen so that eratures are apparent (Chandler 1967, unique air columns were successively Kopec 1973). Weather mobile data also intercepted at two flight levels as they show as much as 2 gm kgm-I and a 1C passed over the urban area. The flight variation in moisture and temperature, pattern thus minimized advection. Hori- respectively, over distances of 1 km. zontal turbulent mixing was found to be Patterns of specific humidity and insignificant and downward entrainment potential temperature are shown for of dry air into the mixing layer was 450 m MSL (300 m AGL) in figure 3. A initially assumed insignificant with the single widespread dry region exists presence of a subsidence inversion. along the southern edge of St. Louis, Because surface data were acquired with tongues of dry air extending along from fixed stations, the time rate of the eastern and western edges of the change of surface moisture measure- metropolitan area. A single heat island ments included advection. Calculations is also apparent and corresponds to the showed that surface moisture advection dry region. Both humidity and temp- on the mesoscale was insignificant. Local erature patterns are reasonably well advection, however, was not assessable correlated with those on the surface. and may be important. The patterns of specific humidity and Temperature and moisture compar- potential temperature patterns at 900 m ison between corrected weather mobiles, MSL (fig. 4) correlate well with those at hygrothermographs, and telemetry sta- 450 m MSL. Both dry regions and heat tions at locations where data from all fields are more widespread at the upper three uniquely different sources could be level. compared, showed consistent results. A From data acquired at the three levels similar consistence was observed be- between 1530 and 1630 CDT, three- tween aircraft and radiosonde measure- dimensional analyses of the moisture ments in the free atmosphere. It is fields in the mixing layer were con- important to recognize that the charac- structed. Isopleth analysis of the specific teristic of the moisture anomalies de- humidity is shown for 10.2 gm kgm-I scribed in this section were consistent (fig. 5) and 11.8 gm kgm-I (fig. 6). A among four independent types of in- plume-like structure with an hourglass struments. configuration is evident. The drier sur- Data from 26 July 1974 for 1530 to face air is apparently mixed vertically 1630 CDT are presented to illustrate through the mixing layer. Measure- iiioisture patterns that have been ob- ments of the wind field convergence also Figure I.-Specific humidity (gm kgm-1) and potential temperature

(OK) patterns for the surface at 1600 CDT. Heavy dashed lines show weather mobile transects. The shaded region defines urban and suburban areas.

0 TELEMETRY STATIONS 26 JULY 74 ILL. STATE WATER SURVEY SURFACE

Figure 2.-St. Louis and surrounding area land use characteristics. Roadways and their corresponding route numbers are shown. The solid triangles represent the St. Louis and Troy navisational vortacs. Figure 3.-Specific humidity (gm kgrn-l) and potential temperature (OK) patterns at 0.45 km MSL (1530-1630 CDT).

Figure 4.-Specific humidity (gm kgm.') and potential temperature (OK) patterns at 0.9 krn MSL (1530-1630 CDT). Figure 5.-Three dimensional analysis of the 10.2 grn kgrn-1 isohume (1530-1630 CDT) for 26 July 1974.

Figure 6.-Three dimensional analysis of the 11.8 gm kgm-1 isohume (1530-1630 CDT) for 26 July 1974. Figure 7.-Upwind aircraft sounding at 1533 CDT. SPECIFIC POTENTIAL m b HUMIDITY TEMPERATURE m b 26 JULY 74 1533 CDT AIRCRAFT (UPWIND)

Fiaurea 8.-Tower Grove Park radiosonde sounding at 1500 CDT. SPECIFIC POTENTIAL m b HUMIDITY TEMPERATURE m b 26 JULY 74 500€ 1500 CDT 1500 PARK

Figure 9.-Downwind aircraft sounding at 1500 CDT. SPECIFIC POTENTIAL HUMIDITY TEMPERATURE m b 26 JULY 74 1500 CDT AIRCRAFT (DOWNWIND) support the concept of convergence and downward mixing across the inversion ascent in the dry regions. will also dry out the columns. Aircraft and radiosonde soundings Overall, the urban-influenced atmo- were also analyzed for the effects of sphere is drier than the surrounding urban modification on vertical moisture rural area. Because there is less moisture profiles. Aircraft soundings were extra- potential at the city surface (vegeta- polated to the surface, using local sur- tion, porous soils, etc.), less energy is face values of humidity and temperature. used for evapotranspiration purposes. Likewise, values from an aspirated psy- The excess energy is then utilized be- chrometer established the surface tem- tween sensible heat energy and ground perature and humidity for radiosondes. stored energy, which generally causes Three soundings (upwind, over, and the urban atmosphere to be hotter. downwind to St. Louis), taken within a In summary of the 26 July 1974 study, period of 30 minutes, are shown in dry regions are found within the St. figures 7 to 9. Particular attention is Louis urban area. These regions of low focused on the lowest 50 mb moisture specific humidity correlate well with im- profile of each sounding. The upwind pervious surfaces found in the densely sounding shows a slight decrease in populated residential areas and light specific humidity with height. The industrial areas. Although the moisture sounding at Tower Grove Park, (5 km content of air columns passing over the southwest of downtown St. Louis), city may decrease from downward en- taken near a large concrete area, shows trainment of dry air from above the a large increase of specific humidity mixing layer, reduced evapotranspira- with height. The downwind sounding tion at the surface clearly contributes to shows only a slight increase of humidity a reduction in average specific humidity with height. These soundings show that of the columns. It can be inferred that within the lowest 50 mb the urban sur- up to 50 percent of the St. Lot~isarea face is drier than the air above 950 mb (Mnrotz and Coiner 1973) is represented on this day. These features are charac- by modified surfaces. Thus the urban teristic of a number of case study days. surface environment is drier due to re- Hence, as air columns pass over the duced evapotranspiration from wide- urban area from a region of high sur- spread impervious surfaces. face humidity (upwind) to a region of lower surface humidity, the mixing of RELEVANCE TO URBAN-DWELLERS dry air into the columns reverses the From energy budget considerations, low-level humidity gradient and de- nearly equal amounts of radiation are creases the average specific humidity of available at both urban and rural sur- the columns. faces (Maisel 1971, Rouse et nl. 1973); These soundings also show that the however, less radiation is used as latent inversion is elevated over the urban area heat energy, so that sensible heat energy by 600 m. The sharp moisture discon- and ground stored energy must be tinuity at the base of the inversion of greater in the urban area. Loose soil the upwind sounding is not present over and plant material characteristically are the city. There is an apparent eroding very poor heat conductors and have a of the air beneath the inversion, prob- low heat capacity. Hence all heat trans- ably attributable to nonuniform down- actions take place in a thin surface ward entrainment of dry air. This layer in most rural areas. In contrast, violates the previous assumption of urban areas have massive surfaces of negligible entrainment into the mixing stone, brick, concrete and asphalt. These layer. This entrainment term could not have better heat conductivity than soils be assessed, but it is apparent that and therefore may be warmed to ap- preciable depth by incoming solar radia- tensity of thunderstorms could be ex- tion. Nocturnally the stored heat is pected to decrease over the urban area, gradually released, which keeps the city as observed by Boatman (1974). This surfaces warmer than surrounding may further result in precipitation grassland or trees. In addition, the anomalies observed downwind of large heated vertical surfaces of buildings urban areas. The small amounts of radiate toward each other and do not water vapor absorbed by hygroscopic lose heat by radiation toward the sky as aerosols generated from industrial ef- a horizontal surface in the open country fluents create haze particles that se- would (Landsberg and Mmisel 1972). verely reduce visibility in the urban The increase in sensible heat energy area. The effect of increased moisture (the heat island) is a well-documented from industry at night when mixing is phenomenon in urban areas. High en- only within a shallow layer would be vironmental temperatures affect health. expected to increase urban humidity, Prolonged exposure to high tempera- such that the city may be more moist tures associated with summer heat than surrounding country air. This may waves can result in death either as a also explain the greater tendency for primary cause or as a contributing fac- fog in urban areas. tor in heart disease, strokes, and pul- Finally, now that evidence for an monary disorders. There is evidence to urban area as a potential weather modi- suggest that a large percentage of heat- fier on the basis of its physical presence related deaths may be of man's own is realized, some recommendations may design by climate modification through be offered that would help to make the urbanization. Clarke and Bach (1 971 ) city a more comfortable place in which showed a relationship between average to work and live. From energy and daily temperature and deaths attributed moisture budget considerations, green primarily to heat during the July 1966 belts or vegetated areas are necessary St. Louis heat wave, in which 500 deaths within our cities. Green belts would occurred in East St. Louis, Illinois. Al- not only help to embellish the downtown though no heat deaths occurred at an image, but also moderate the cities average temperature of 31.7C, deaths temperature. Large asphalt parking increased markedly with every one- lots should be transected with trees and degree increase in temperature above shrubs to reduce overheating. Fountains 32.2C (assumed the critical level). A and small parks should be constructed rise of the average daily temperature of within the heart of the city where the only 0.6C above 32.2C resulted in 11 greatest percentage of concrete and heat deaths. Seventy-three heat deaths impervious surfaces exist. Gardens and occurred when the average daily tem- vegetation could be planted on rooftops. perature was 35C. This 2.8C increment All of these modifications of the urban of average daily temperature is about moisture budget would contribute to a the same as the average temperature moderation of the daytime temperature difference between the urban and sub- excess of the urban heat island. urban environments, based on the after- noon and evening temperatures obtained CONCLUSIONS in Clarke and Bach's study. The lack of available moisture from Other effects of the urban moisture the urban surface leads to increased climate may also contribute to man's dis- sensible heat energy and ground stored comfort in his atmospheric environment. energy. Properly spaced green areas The reduced moisture content of urban- are perhaps the most important means influenced air may locally deprive storms of controlling the urban temperature of latent heat energy, so that the in- excess. The urban heat island effect could also be reduced in torrid latitudes Kopec, R. J. 1973. DAILYSPATIAL AND SECULAR VARIATIONS using that have OF ATMOSPHERIC HUMIDITY IN A SMALL CITY. lower heat conductivity and capacity. J. Apple Meteorol. 12: 639-6480 Landsberg, H. E., and T. N. Maisel. with different and 1972. MICROMETEOROLOGICALOBSERVATIONS IN colors, through changes in albedo and AN AREA OF URBAN GROWTH. Boundarv-La~er Meteorol. 12 : 365-370. heat absorption, would cause differen- ~~i~~l,T. N. in heating and encourage mixing 1971. EARLYMICROMETEOROLOGICAL CHANGES CAUSED BY URBANIZATION. M.S. thesis, Inst. and ventilation. On the Fluid Dynam. and Appl. Math., Univ. Md. cities in high latitudes may be purpose- College Park. 50 P. Marotz, G. A., and J. C. Coiner. strengthen the heat- 1973. ACQUISITIONAND CHARACTERIZATION OF island effect. If the effects of urban SURFACE MATERIAL DATA FOR URBAN CLI- MATOLOGICAL STUDIES. J. Appl. Meteorol. 12: development on the local climate are 919-923. considered in the urban planning pro- MY~UP,L. 0. 1969. A NUMERICAL MODEL OF THE URBAN cess, climate modification can be pur- HEAT ISLAND. J. Appl. Meteovol. 8: 908-918. posely controlled toward man's comfort Rouse, W. R., D. Noad, and J. lbfcchtcheon. 1973. RADIATION, TEMPERATURE AND AT- rather than to MOSPHERIC EMISSIVITIES IN A POLLUTED develop to man's distress. URBAN ATMOSPHERE AT HAMILTON,ONTARIO. J. Appl. Meteorol. 12: 798-807. Sisterson, D. L. 1975. STUDIES ON THE URBAN MOISTURE LITERATURE CITED BUDGET. Univ. Wvo. Coil. Enz. Den- Atmos. Sci. Rep. AS ll4."~aramie.53-p. Boatman, J. F. Terjung, W. H., G. E. Borgel, R. N. Hawkins, 1974. INADVERTENTTHUNDERSTORM MODIFICA- R. D. Hickey, R. B. Howard, R. N. Kickert, TION BY AN URBAN AREA. M.S. thesis, Dep. S. Louis, G. L. Potter, S. E. Tuller, and T. Van Atmos. Resour., Coll. Eng., Univ. Wyo. Hueklon. Laramie. 126 p. 1971. THE EFFECT OF A CYCLONIC STORM ON Chandler, T. J. THE ENERGY FLUXES AT THE URBAN INTERFACE 1967. ABSOLUTEAND RELATIVE HUMIDITY OF -A PRELIMINARY EXPERIMENT. Arehiv. fiir TOWNS. Am. Meteorol. Soe. Bull. 48: 394-399. Il.leteoro1. Geovhus.- - u. Bioklimatol. Ser. B, Clarke, J. F., and W. Bach. 19: 367-416. 1971. COMPARISONOF THE COMFORT CONDI- Wood, J. L. TIONS IN DIFFERENT URBAN AND SUBURBAN 1971. THE NOCTURNAL HEAT ISLAND IN MICROENVIRONMENTS. znt. J. Biometeorol. AUSTIN,TEXAS. Univ. Tex. Coil. Eng. Atmos. 15: 41-54. Sci. Group. Rep. 28, Austin. 55 p.