• Bornstein, R., and Q. Lin. 2000. Urban heat islands and summertime convective thunderstorm in Atlanta: Three cases studies. Atmos. Environ. 34:507516.

    • Search Google Scholar
    • Export Citation
  • Comarazamy, D. E., and J. E. González. 2008. On the validation of the simulation of early season precipitation in the island of Puerto Rico using a mesoscale atmospheric model. J. Hydrometeor. 9:507520.

    • Search Google Scholar
    • Export Citation
  • Cotton, W. R. Coauthors 2003. RAMS 2001: Current status and future directions. Meteor. Atmos. Phys. 82:529.

  • Daly, C., , E. H. Helmer, , and M. Quiñones. 2003. Mapping the climate of Puerto Rico, Vieques and Culebra. Int. J. Climatol. 23:13591381.

    • Search Google Scholar
    • Export Citation
  • Dickinson, R. E., , A. H. Sellers, , P. J. Kennedy, , and M. F. Wilson. 1986. Biosphere-Atmosphere Transfer Scheme (BATS) for the NCAR Climate Community Model. NCAR Tech. Note NCAR/TN-275+STR. 69 pp.

    • Search Google Scholar
    • Export Citation
  • Dixon, P. G., and T. L. Mote. 2003. Patterns and causes of Atlanta’s urban heat island–initiated precipitation. J. Appl. Meteor. 42:12731284.

    • Search Google Scholar
    • Export Citation
  • González, J. E. Coauthors 2005. Urban heat islands developing in coastal tropical cities. Eos, Trans. Amer. Geophys. Union 86.doi:10.1029/2005EO420001.

    • Search Google Scholar
    • Export Citation
  • González, J. E., , J. C. Luvall, , D. Rickman, , D. E. Comarazamy, , and A. J. Picón. 2006. Urban heat island identification and climatological analysis in a coastal, tropical city: San Juan, Puerto Rico. Urban Remote Sensing, Q. Weng and D. Quattrochi, Eds., CRC Press, 223–252.

    • Search Google Scholar
    • Export Citation
  • Inoue, T., and F. Kimura. 2004. Urban effects on low-level clouds around the Tokyo metropolitan area on clear summer days. Geophys. Res. Lett. 31:L05103. doi:10.1029/2003GL018908.

    • Search Google Scholar
    • Export Citation
  • Jauregui, E. 1997. Heat island development in Mexico City. Atmos. Environ. 31:38213831.

  • Jauregui, E., and E. Romales. 1996. Urban effects of convective precipitation in Mexico City. Atmos. Environ. 30:33833389.

  • Kalnay, E. Coauthors 1996. The NCEP/NCAR 40-Year Reanalysis Project. Bull. Amer. Meteor. Soc. 77:437471.

  • Landsberg, H. E. 1981. The Urban Climate. Academic Press, 275 pp.

  • Lawton, R. O., , U. S. Nair, , R. A. Pielke Sr., , and R. M. Welch. 2001. Climatic impact of tropical lowland deforestation on nearby montane cloud forests. Science 294:584587.

    • Search Google Scholar
    • Export Citation
  • Luvall, J. C., and H. R. Holbo. 1989. Measurement of short-term thermal responses of coniferous forest canopies using thermal scanner data. Remote Sens. Environ. 27:110.

    • Search Google Scholar
    • Export Citation
  • Luvall, J. C., , D. Lieberman, , M. Lieberman, , G. S. Hartshorn, , and P. Peralta. 1990. Estimation of tropical forest canopy temperatures, thermal response numbers, and evapotranspiration using an aircraft-based thermal sensor. Photogramm. Eng. Remote Sens. 56:13931401.

    • Search Google Scholar
    • Export Citation
  • Meyers, M. P., , R. L. Walko, , J. Y. Harrington, , and W. R. Cotton. 1997. New RAMS cloud microphysics parameterization. Part II: The two-moment scheme. Atmos. Res. 45:339.

    • Search Google Scholar
    • Export Citation
  • Nair, U. S., , R. O. Lawton, , R. M. Welch, , and R. A. Pielke Sr.. 2003. Impact of land use on Costa Rican tropical montane cloud forests: Sensitivity of cumulus cloud field characteristics to lowland deforestation. J. Geophys. Res. 108:4206. doi:10.1029/2001JD001135.

    • Search Google Scholar
    • Export Citation
  • Noto, K. 1996. Dependence of heat island phenomena on stable stratification and heat quantity in calm environment. Atmos. Environ. 30:475485.

    • Search Google Scholar
    • Export Citation
  • Oke, T. R. 1987. Boundary Layer Climates. Cambridge University Press, 450 pp.

  • Pielke, R. A. Coauthors 1992. A comprehensive meteorological modeling system—RAMS. Meteor. Atmos. Phys. 49:6991.

  • Poreh, M. 1996. Investigation of heat islands using small scale models. Atmos. Environ. 30:467474.

  • Ray, D. K., , U. S. Nair, , R. O. Lawton, , R. M. Welch, , and R. A. Pielke Sr.. 2006. Impact of land use on Costa Rican tropical montane cloud forests: Sensitivity of orographic cloud formation to deforestation in the plains. J. Geophys. Res. 111:D02108. doi:10.1029/2005JD006096.

    • Search Google Scholar
    • Export Citation
  • Rickman, D. L., , J. C. Luvall, , and S. Schiller. 2000. An algorithm to atmospherically correct visible and thermal airborne imagery. Workshop on Multi/Hyperspectral Sensors, Measurements, Modeling, and Simulation, Huntsville, Alabama, U.S. Army Aviation & Missile Command, SR-RD-WS-01-01.

    • Search Google Scholar
    • Export Citation
  • Rozoff, C. M., , W. R. Cotton, , and J. O. Odegoke. 2003. Simulation of St. Louis, Missouri, land use impacts on thunderstorms. J. Appl. Meteor. 42:716738.

    • Search Google Scholar
    • Export Citation
  • Saleeby, M. S., and W. R. Cotton. 2004. A large-droplet mode and prognostic number concentration of cloud droplets in the Colorado State University Regional Atmospheric Modeling System (RAMS). Part I: Module descriptions and supercell test simulations. J. Appl. Meteor. 43:182195.

    • Search Google Scholar
    • Export Citation
  • Shepherd, J. M. 2005. A review of current investigations of urban-induced rainfall and recommendations for the future. Earth Interactions 9. [Available online at http://EarthInteractions.org].

    • Search Google Scholar
    • Export Citation
  • Shepherd, J. M., and S. J. Burian. 2003. Detection of urban-induced rainfall anomalies in a major coastal city. Earth Interactions 7. [Available online at http://EarthInteractions.org].

    • Search Google Scholar
    • Export Citation
  • Taylor, M. A., , D. B. Enfield, , and A. A. Chen. 2002. Influence of the tropical Atlantic versus the tropical Pacific on Caribbean rainfall. J. Geophys. Res. 107:3127. doi:10.1029/2001JC001097.

    • Search Google Scholar
    • Export Citation
  • Tso, C. P. 1996. A survey of urban heat island studies in two tropical cities. Atmos. Environ. 30:507519.

  • Van der Molen, M. K. 2002. Meteorological impacts of land use change in the maritime tropic. Ph.D. thesis, Vrije Universititeit Amsterdam, 262 pp.

  • Van der Molen, M. K., , A. J. Dolman, , M. J. Waterloo, , and L. A. Bruijnzeel. 2006. Climate is affected more by maritime then by continental land use change: A multiple scale analysis. Global Planet. Change 54:128149.

    • Search Google Scholar
    • Export Citation
  • Velazquez-Lozada, A., , J. E. Gonzalez, , and A. Winter. 2006. Urban heat island effect analysis in San Juan, Puerto Rico. Atmos. Environ. 40:17311741.

    • Search Google Scholar
    • Export Citation
  • View in gallery

    Temperature profile observed in the SJMA, based on data collected during the ATLAS Mission (February 2004) from surface weather stations and temperature sensors located along the east–west solid line in the bottom panel of Figure 3 and constructed with the average daytime temperatures.

  • View in gallery

    Model grids used in the numerical simulations. The topography contours have intervals of 150 m.

  • View in gallery

    Specification of the surface characteristics used in the two atmospheric model runs used for the study for the (top) natural and (bottom) urban scenarios. The reader is referred to Table 1 for a description of each class number and its corresponding biophysical parameters. The solid line in the urban scenario represents the transect line where the stations and sensors used to construct Figure 1 are located.

  • View in gallery

    RAMS/ATLAS overall data processing flow to produce atmospheric, radiometric, and geometrically corrected data and link to update the mesoscale model surface characteristics (adapted from Rickman et al. 2000).

  • View in gallery

    ATLAS sensor visible bands 1–3: this information was used to identify the extension and morphology of the SJMA.

  • View in gallery

    ATLAS-derived albedo dataset used to configure the surface characteristics of RAMS.

  • View in gallery

    Comparison of canopy temperatures between those simulated by RAMS and those observed by the ATLAS sensor when it was flown over the SJMA.

  • View in gallery

    Comparison of the air temperatures between the regional model results and the stations and sensors deployed in the SJMA.

  • View in gallery

    Vertical profiles of temperature (°C) and wind speed (m s−1) for model results [solid line is SJMA ATLAS; dashed line is SJMA homogeneous slab] and the San Juan NWS office balloon data (open squares) for (a) 1200 UTC 11 Feb, (b) 0000 UTC 14 Feb, and (c) 1200 UTC 16 Feb 2004.

  • View in gallery

    Spatial distribution of the air temperature difference (°C) at 2 m AGL between the urban and natural scenarios simulated for the analysis.

  • View in gallery

    (left) Wind field averaged at 1500 LST during the complete period of simulation of the present run. (right) Model wind field difference between the urban and natural runs averaged at 1500 LST.

  • View in gallery

    (left) Total accumulated precipitation for the urban run during the ATLAS period (in mm and limited to the 9-mm contour because the El Yunque total, >50 mm, masked precipitation in other regions) and (right) precipitation difference (in percentage increase) between the urban and natural runs.

  • View in gallery

    Increased sensible heat flux over the urban area for the urban and natural scenario simulations.

  • View in gallery

    Simulated vertical motion fields (10−1 m s−1) during 14 Feb 2004 at the specified times for the (top) urban and (bottom) natural runs. Early morning hour panels were omitted because of the suppressed vertical motions that are typical for calm and stable conditions at this time of the day.

  • View in gallery

    Fractional cloud coverage predicted by RAMS for the urban run at selected times on 14 Feb 2004.

  • View in gallery

    North–south vertical cross sections of wind vectors difference between the urban and natural runs for the average (top left)–(top right) 1200, 1400, and 1600 LST and (bottom left)–(bottom right) 1800, 2000, and 2200 LST. The thick horizontal line at the bottom of each panel depicts the presence of land, and the two thick vertical lines represent the two reference topographic peaks of ∼700 and 200 m.

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 12 12 6
PDF Downloads 6 6 0

A Land–Atmospheric Interaction Study in the Coastal Tropical City of San Juan, Puerto Rico

View More View Less
  • 1 Department of Mechanical Engineering, Santa Clara University, Santa Clara, California
  • 2 Department of Mechanical Engineering, City College of New York, New York, New York
  • 3 Global Hydrology and Climate Center, NASA Marshall Space Flight Center, Huntsville, Alabama
  • 4 Iberdrola Renewables, Portland, Oregon
© Get Permissions
Full access

Abstract

This paper focuses on the surface–atmospheric interaction in a tropical coastal city including the validation of an atmospheric modeling and an impact study of land-cover and land-use (LCLU) changes. The Regional Atmospheric Modeling System (RAMS), driven with regional reanalysis data for a 10-day simulation, is used to perform the study in the San Juan metropolitan area (SJMA), one of the largest urban conglomerations in the Caribbean, which is located in the island of Puerto Rico and taken as the test case. The model’s surface characteristics were updated using airborne high-resolution remote sensing information to obtain a more accurate and detailed configuration of the SJMA. Surface and rawinsonde data from the San Juan Airborne Thermal and Land Applications Sensor (ATLAS) Mission are used to validate the modeling system, yielding satisfactory results in surface/canopy temperature, near-surface air temperatures, and vertical profiles. The impact analysis, performed with the updated SJMA configuration and a potential natural vegetation (PNV) scenario, showed that the simulation with specified urban LCLU indexes in the bottom boundary produced higher air temperatures over the area occupied by the city, with positive values of up to 2.5°C. The same analysis showed changes in the surface radiative balance in the urban case attributed to modifications in the LCLU. This additional heat seems to motivate additional vertical convection that may be leading to possible urban-induced precipitation downwind of the SJMA. This was evident in a precipitation disturbance when the city is present (∼0.9 mm, 22.5% increase) captured by the model that was accompanied by increases in cloud formation and vertical motions mainly downwind of the city.

* Corresponding author address: Jorge E. González, Department of Mechanical Engineering, The City College of New York, Steinman Hall (T-238), New York, NY 10031. gonzalez@me.ccny.cuny.edu

Abstract

This paper focuses on the surface–atmospheric interaction in a tropical coastal city including the validation of an atmospheric modeling and an impact study of land-cover and land-use (LCLU) changes. The Regional Atmospheric Modeling System (RAMS), driven with regional reanalysis data for a 10-day simulation, is used to perform the study in the San Juan metropolitan area (SJMA), one of the largest urban conglomerations in the Caribbean, which is located in the island of Puerto Rico and taken as the test case. The model’s surface characteristics were updated using airborne high-resolution remote sensing information to obtain a more accurate and detailed configuration of the SJMA. Surface and rawinsonde data from the San Juan Airborne Thermal and Land Applications Sensor (ATLAS) Mission are used to validate the modeling system, yielding satisfactory results in surface/canopy temperature, near-surface air temperatures, and vertical profiles. The impact analysis, performed with the updated SJMA configuration and a potential natural vegetation (PNV) scenario, showed that the simulation with specified urban LCLU indexes in the bottom boundary produced higher air temperatures over the area occupied by the city, with positive values of up to 2.5°C. The same analysis showed changes in the surface radiative balance in the urban case attributed to modifications in the LCLU. This additional heat seems to motivate additional vertical convection that may be leading to possible urban-induced precipitation downwind of the SJMA. This was evident in a precipitation disturbance when the city is present (∼0.9 mm, 22.5% increase) captured by the model that was accompanied by increases in cloud formation and vertical motions mainly downwind of the city.

* Corresponding author address: Jorge E. González, Department of Mechanical Engineering, The City College of New York, Steinman Hall (T-238), New York, NY 10031. gonzalez@me.ccny.cuny.edu

1. Introduction

Human activity in urban environments has impacts in the regional scale, such as changing the atmospheric composition, affecting the water cycle, and modifying ecosystems. Nevertheless, our understanding of the role of urbanization in the ecoclimate system is incomplete, yet it is critical to determine how in coastal environments the atmosphere–ocean–land–biosphere components act reciprocally in a connected system. The most clear mesoscale indicator of environmental impacts due to urbanization is a well-known urban/rural convective circulation known as urban heat islands (UHIs). The urban heat island is defined as a dome of high temperatures observed over urban centers as compared to the relatively low temperatures of the rural surroundings (see Figure 1).

Some of the factors that lead to the formation of a heat island are the widespread use of diverse construction materials. Materials used in cities have a much higher thermal inertia than natural vegetation-covered surfaces, resulting in large differences in temperature during the first several hours after sundown, when all the energy absorbed and stored during daylight is released to the lower atmosphere over the city. Tall vertical surfaces and other geometric shapes of the urban landscape create what is known as the canyon effect, another aspect of the UHI effect. In the spaces between buildings, longwave radiation emitted by the surface at night is reflected and absorbed by the walls, resulting in trapped energy and higher temperatures. The urban topography also interrupts wind flows and results in decreased heat loss. Nonporous paved surfaces prevent precipitation from entering the soil and subsoil, limiting evapotranspiration in urban/built environments, a natural phenomenon with an evaporative cooling effect. These temperature contrasts are greater in clear and calm conditions and tend to disappear in cloudy and windy weather by effects of thermal and mechanical mixing.

UHI effects of diverse magnitude have been reported for a number of cities (Landsberg 1981; Tso 1996; Jauregui 1997; Noto 1996; Poreh 1996; Rozoff et al. 2003; Dixon and Mote 2003; Shepherd 2005). The majority of these studies focus on large continental cities, generally located in temperate northern latitudes; although the general pattern is very similar, each city is exposed to diverse local and synoptic factors, which causes the study of UHI to be complex and specific of the locality.

Tropical coastal areas represent an interesting case in which global, regional, and local effects converge. However, studies of land-cover and land-use (LCLU) changes in these locations have been very limited. It was recently reported in a series of atmospheric modeling studies that low-land deforestation is leading to increases in cloud-base heights and thinner clouds in rain forests in Central America (Lawton et al. 2001; Nair et al. 2003; Ray et al. 2006), which is resulting in increases of regional droughts. However, similar numerical experiments conducted in Puerto Rico reported apparently contradictory results where forested runs produced higher cloud bases when compared to pasture simulations (Van der Molen 2002; Van der Molen et al. 2006). The discrepancy on the results by these two studies is settled by explaining that the effect on cloud-base height is linked to the net effect the LCLU change has on the local Bowen ratio (Ray et al. 2006). It was argued that the Puerto Rican pasture fields were specified as well-watered sites, even during the early rainfall season (ERS), and the forest was made to be coastal swamp forests, which results in the forests having less evaporation and higher sensible heat fluxes than the well-watered pastures. In the Costa Rican studies, on the other hand, the forests were specified with constant access to deep soil water, whereas pastures were not, resulting in the reported increase in Bowen ratio values in the deforestation cases. Because warm coastal UHI effects may induce precipitation (Shepherd and Burian 2003; Shepherd 2005), the reported increase in cloud-base heights due to deforestation produces net competing effects not investigated before.

Induced urban precipitation may be attributed to enhanced sensible heating of air directly in contact with the increasingly urbanized or deforested surface, which may lead to further convective destabilization of low-level air. This effect, combined with the surface convergence associated with increased surface roughness, particularly in heavily urbanized settings, leads to stronger updrafts. This would favor the formation of convective clouds and lead to increased precipitation over and downwind of these areas. On the other hand, the reduction of surface moisture that occurs as a result of a hampered evapotranspiration process, associated with the removal of water-rich biosphere, raises the lifting condensation level (LCL). In other words, drier surface conditions yield higher cloud bases, as represented by the LCL, which will make it more difficult for moist convection to organize and for condensates to fall out and reach the ground.

Velazquez-Lozada et al. (Velazquez-Lozada et al. 2006) conducted an observational and numerical study that proved the existence UHI for the tropical coastal city of San Juan, Puerto Rico, and some of the potential impacts of increasing urbanization on the local climate. They found that UHI conditions (larger temperature differences between the city and the surrounding countryside) were more noticeable during the Caribbean dry and early rainfall seasons, which together compose the period from December to June, when weather is drier and calmer at both synoptic and convective scales. The late rainfall season, from September to November, is more humid and convectively active (Daly et al. 2003; Taylor et al. 2002).

It may be assumed that, with such a large and constant supplier of water vapor like the ocean via synoptic-scale trade wind advection, coastal tropical cities will not be affected by drought-inducing higher cloud bases due to changes in the Bowen ratio over land. However, this hypothesis remains inconclusive, and we believe it underestimates the complexity of the competing multiscale factors involved in these land–ocean–atmosphere interactions. Thus, a number of research questions arise regarding the understanding of these and other competing effects due to LCLU changes in tropical coastal regions. These effects also include the feedbacks and interactions associated with possible changes in the hydrological cycle such as convection, evapotranspiration, condensation, cloud coverage, precipitation, and runoff and those associated with changes in the interaction of the sea-breeze fronts with large coastal UHIs. Geographical areas where these common factors can be encountered are those regions where the mass of land is relatively small with respect to the surrounding sea, such as Central America and the Caribbean Islands and regions and islands in Southeast Asia.

These questions motivated the development of more extensive and wide-ranging experimental campaigns and numerical experiments with the objective of determining the particular characteristics of the UHI in San Juan and investigating the possible impacts of LCLU changes in tropical coastal regions. The first of these field campaigns was designated as the San Juan Airborne Thermal and Land Applications Sensor (ATLAS) Mission, which took place on 11–16 February 2004. More details of the San Juan ATLAS Mission can be found in González et al. (González et al. 2005; González et al. 2006). The observations obtained during this mission are used in this paper in two ways. First, the airborne remote sensing information is used to identify the geographical extension of the San Juan metropolitan area (SJMA) and to update some of the model’s out-of-the-box surface characteristics. Second, surface and rawinsonde data are used to validate the regional atmospheric model chosen for the study with the updated urban LCLU configuration. The objective is to obtain a detailed, heterogeneous configuration of the SJMA that allows performing a climate impact analysis from LCLU changes. Both of these aspects of the study are expanded in subsequent sections. With this approach, we also report climatic impacts of low-land urbanization in this tropical coastal region. The SJMA represents an interesting case study 1) because it is one of the largest urban areas in the Caribbean with an extremely high concrete density but without very tall buildings or high rises, which helps with the methodology presented in this paper; 2) because its close proximity to a tropical montane cloud forest (El Yunque) and the Central Mountains range makes the quantification of climate changes an important scientific and social matter; and 3) because it is a region where global, regional, and local atmospheric circulations converge. Although these conditions are also present in other tropical cities, the availability of the remote sensing data and a large historical climate dataset make the SJMA an ideal test case.

The main objective of the numerical simulations presented herein is to investigate the feasibility and reliability of a regional atmospheric modeling system to be used in studies of the impact of land usage for urbanization on different environmental variables at local and regional scales in tropical coastal areas. The approach used includes the configuration of a mesoscale atmospheric model using airborne remote sensing information, validation of the control simulations with the ATLAS Mission observations, and a preliminary test of the impact of LCLU changes.

2. Numerical experiments

The model chosen for the study is the Regional Atmospheric Modeling System (RAMS), a highly versatile numerical code developed to simulate and forecast meteorological phenomena at Colorado State University (Pielke et al. 1992; Cotton et al. 2003). The version of RAMS used in this investigation, a modification of the standard v4.3 release, contains an upgraded cloud microphysics module described by Saleeby and Cotton (Saleeby and Cotton 2004), an advancement of the original package available in the current model release (Meyers et al. 1997).

2.1. Environmental setting

Soundings from several locations across Puerto Rico, taken at various times during the ATLAS Mission, show a strong 5°–10°C temperature inversion anywhere within the 800–750-mb layer (the so-called trade wind inversion; see Figure 8 for the San Juan temperature and wind sounding data). Considerable vertical wind shear at mid- to upper levels is also observed as well as relatively dry surface conditions leading to dry and stable state of the atmosphere over the Caribbean region, ideal conditions for conducting urban heat island studies (Velazquez-Lozada et al. 2006). Moreover, these are the conditions during which the atmospheric model used in this study has been proven to perform satisfactorily for monthly precipitation prediction over the island of Puerto Rico (Comarazamy and González 2008). The general atmospheric conditions described above prevailed during the 6-day duration of the mission.

2.2. General model configuration

The simulations were conducted with three grids making use of the grid nesting capabilities of the model. Grid 1 covers a great part of the Caribbean basin with a horizontal resolution of 25 km. Grid 2, which is nested within grid 1, covers the island of Puerto Rico with a horizontal resolution of 5 km. Grid 3 is nested within grid 2 and centered in the city of San Juan with a resolution of 1 km (see Figure 2). For the vertical coordinate, all the grids have the same specification. A vertical grid spacing of 100 m was used near the surface and stretched at a constant ratio of 1.1 until a Δz of 1000 m is reached. The depth of the model is approximately 22.83 km with 40 vertical levels. All the runs performed had a simulation time of 10 days, from 10 to 20 February 2004, and were driven by the same large-scale atmospheric 4D fields as provided by the National Centers for Environmental Prediction (NCEP)–National Center for Atmospheric Research (NCAR) reanalysis data (Kalnay et al. 1996). The horizontal diffusion coefficients are computed based on the original Smagorinsky formulation and vertical diffusion is parameterized with the Mellor and Yamada closure scheme, which employs a prognostic turbulent kinetic energy (TKE) equation. On grid 3, the explicit microphysics scheme of Saleeby and Cotton (Saleeby and Cotton 2004) is activated.

To try to quantify the impact of the LCLU change in the SJMA during these short model simulations, two different scenarios were configured. First, one of the model subroutines was modified to represent the urban extension, configuration, and optical characteristics of San Juan as it was observed from airborne remote sensing data (details are given in next subsection). The second configuration was designed to represent the potential natural vegetation (PNV) of the zone occupied by the city by interpolating the surrounding vegetation to cover the entire area. The runs were denominated urban and natural, respectively. The variable modified for these numerical simulations was the denominated vegetation index (VI) defined by the biosphere–atmosphere transfer scheme (BATS; Dickinson et al. 1986); this index includes the physical parameters of albedo, emissivity, leaf area index (LAI), vegetation percentage, surface roughness, and root depth. The configuration of the LCLU index used in the two simulations is presented in Figure 3.

2.3. ATLAS sensor and data processing

The ATLAS operates in the visual and infrared bands. It can detect 15 multispectral channels of the radiation through the visible, near-infrared, and thermal spectrums. The data are corrected for atmospheric radiation and georectified before any analysis is performed. The ATLAS sensor was flown over San Juan, Puerto Rico, during February of 2004 to investigate the impact of the urban growth and landscape in the climate of this tropical city. For more details on the ATLAS sensor and the San Juan ATLAS Mission, please refer to González et al. (González et al. 2005; González et al. 2006).

To calibrate the ATLAS data, detailed atmospheric corrections are needed; these require direct measurements of the atmosphere extinction coefficients by wavelength, ATLAS instrument characteristics and calibration, and profiles of atmospheric temperatures and water vapor (provided by weather balloon launched during the mission). A combination of commercial and public domain software was used for the processing, including the series of custom programs Watts and Energy from Earth Resources Laboratory Applications Software V.II (ELASII; Rickman et al. 2000). Rickman et al. (Rickman et al. 2000) detail the procedure for calibrating the ATLAS sensor to produce the system transfer function to convert digital values (DVs) into radiance measurements. These procedures produce ATLAS data files that are in physical units of energy, which are used to generate files that derive albedo and surface temperature. Figure 4 shows the process flowchart followed for this project, including the link between the airborne remote sensing and the mesoscale atmospheric modeling.

Surface temperature is a major component of the surface energy budget. Use of energy terms in modeling surface energy budgets allows the direct comparison of various land surfaces encountered in a landscape from vegetated (forest and herbaceous) to nonvegetated (bare soil, roads, and buildings; Oke 1987). The partitioning of energy budget terms depends on the surface type. In natural landscapes, the partitioning is dependent on canopy biomass, leaf area index, aerodynamic roughness, and moisture status, all of which are influenced by the development stage of the ecosystem. In urban landscapes, coverage by man-made materials substantially alters the surface energy budget. The canopy net all-wave radiation balance (W m−2), temperature (°C), and albedo can be determined following Oke (Oke 1987). Because the visible bands and the albedo information from the ATLAS sensor were used extensively for updating the surface characteristics of the atmospheric model (detailed in the next section), a brief description of the formulation to calculate the surface albedo is given.

The net solar radiation S* at the given wavelength is given by
i1087-3562-14-16-1-eq1
where α is the site albedo and S↓ is the incoming solar radiation.
The albedo is defined as
i1087-3562-14-16-1-eq2
where S↑ is the reflected solar radiation.

2.4. Configuration of the SJMA urban surface characteristics

As mentioned before, airborne remote sensing data were used to configure the “urban and built up” vegetation index used for the urban run. This process was the culmination of several steps (Figure 4). To start, a 1 km × 1 km subset of the 10-m ATLAS data was extracted to match the horizontal resolution of the model grid where the analysis is performed. Then the geographical extension of the SJMA was identified using bands 1–3 of the sensor’s visible spectrum, and it is estimated to have a total surface area of 435 km2, as seen in Figure 5 (where the complete 10-m mosaic of the San Juan flight lines is presented). Following this information, the 1 km × 1 km dataset of surface albedos corresponding to the SJMA was obtained using band 1 of the ATLAS sensor thermal spectrum. Figure 6 shows the 10-m mosaic of the ATLAS-derived albedos obtained for the San Juan flight lines.

For practical reasons, it was decided to group the SJMA albedos into five groups: albedos <0.15, 0.16–0.20, 0.21–0.25, 0.26–0.30, and >0.31. Each middle group was given a single albedo value calculated as the average of the range of each one, and the end groups were given a value of ±0.05, respectively, from their adjacent group. This process of averaging the albedos is justified, not only to facilitate the link with the atmospheric model grid but also because the contribution of close albedo values is expected to be similar, whereas the averaged albedos produced a level of heterogeneity that satisfied the authors and still the effect on temperatures was evident (as it will be shown in the next section). Furthermore, the methodology used opens the door for subsequent improvements and additions to the vegetation index to run hypothetical scenarios and to perform simulations that test possible mitigation/adaptation strategies.

Once the SJMA albedo groups were generated, five new vegetation indexes were created following BATS by combining the SJMA albedos with the biophysical parameters of the bare soil [leaf area index, vegetation fraction (VFRAC)] and urban/built-up (surface roughness, vertical displacement) indices. These new VIs were denominated urban 1–urban 5. Table 1 contains the full list of VIs and their corresponding parameters. Finally, to accurately match the location of each of the 1 km × 1 km urban indexes with the atmospheric model grid, a natural landmark was identified and specified in RAMS (San Juan Bay in this case). The rest of the dataset was located to obtain the urban surface configuration shown in Figure 3.

3. Results

3.1. Model validation: Comparison with observations

The model results produced by the urban run were compared with the observations obtained during the ATLAS Mission experimental campaign conducted in Puerto Rico during the month of February 2004 (González et al. 2005). These observational data consist of the canopy temperature of the SJMA and surrounding areas gathered by the ATLAS sensor during the 16 February 2004 flight, air temperatures at 2 m above ground level (AGL) as recorded by weather stations and temperature sensors, and the vertical profiles of temperature and wind speed from weather balloons launchings performed regularly by the San Juan National Weather Service (NWS) office.

As with the ATLAS visible bands and albedo information described in the previous section, the first step in analyzing the sensor canopy temperature values was the extraction of a 1-km horizontal resolution subset. Then the model grid was resized to match the area covered by the plane flight lines; this way, the model’s simulated canopy/vegetation top temperatures were compared with the observed values (Figure 7). Through this exercise, it is shown that, although the 1 km × 1 km averaging of the ATLAS values affects the visual representation of the data, the model was able to reproduce the same spatial pattern of surface heating as the one observed over the SJMA on 14 February 2004 at the time of the flight during the ATLAS Mission. Furthermore, the areas of maximum heating are located in the western sector of the SJMA and the Caguas area in both panels, and both show the same range in temperature values (∼14°–55°C). These results are considered satisfactory, especially when analyzing the way the lower atmosphere interacts with this surface heating in terms of its vertical structure and the daily temperature cycle.

The daily temperature cycle presented in Figure 8 was obtained by averaging the temperature values predicted by the model over the entire area represented by the city at each hour for the duration of the ATLAS Mission, and it was compared with the stations and ground sensors averaged over the same geographical area and time span (10–20 February 2004). This comparison shows that the model performs satisfactorily, even though it produced temperatures slightly higher than the observed ones during the late morning and early afternoon hours when surface heating is more intense. A preliminary configuration using a homogeneous slab of concrete in the area of the city is also shown in Figure 8. The overprediction of this preliminary configuration could be explained by the use of a homogenous urban LCLU; therefore, it is not capturing the different microclimates present in the metropolitan area and producing a more uniform temperature distribution throughout the area.

The improvement in the prediction by using five urban categories ingesting ATLAS-derived albedo information is also evident from Figure 8. Although the model is still has a tendency for overpredicting temperatures from late morning to early afternoon, the results show potential for further improvement by incorporating other in situ measurements of physical parameters that are part of the vegetation index. The reduction of daytime high temperatures in the simulations with more urban details might be due to a reduction of sensible heat flux and an increase in soil heat flux, or heat storage. Because the atmospheric mesoscale model used does not account for the thermal storage of the urban area, a feasible approach would be to use the thermal response number (TRN) developed by Luvall and Holbo (Luvall and Holbo 1989) to characterize the thermal response of different surface types. The TRN is a surface property defined as the ratio of the surface net radiation, which integrates the effects of the nonradiative fluxes over short periods of time, and the rate of change in surface temperature that expresses how those fluxes are reacting to radiant energy inputs (Luvall and Holbo 1989; Luvall et al. 1990).

Vertical profiles of model-resolved temperature and wind speed were compared with sounding data from the San Juan NWS office launched on the dates when the San Juan ATLAS Mission flights took place (Figure 9). It was found that the model follows the general pattern of the observations, especially in the case of the temperature profiles. In the case of the wind speeds, the model performs very well at mid- to high altitudes, capturing the upper-level jet present at approximately 14.5 km. Near the surface (below 2 km), however, the model follows the general trend of the sounding but is unable to capture details of small fluctuations in wind speeds, which might be due to the relatively coarse resolution specified in the vertical coordinate (100 m for the lowest model layer).

3.2. Impact of LCLU changes

To study the impact of LCLU changes over the SJMA and surrounding areas, an analysis of air temperatures at 2 m AGL was performed using the results produced by the urban and natural scenarios. The analysis consists of calculating the difference between the temperature values averaged throughout the simulation period at the time of maximum heating, determined to be 1500 LST [Atlantic standard time (AST)], by performing the simple urban minus rural operation (Figure 10). Here it is shown that the atmospheric model predicts that the presence of San Juan has a strong impact in the lower atmosphere of the area occupied by the city. This impact is reflected in higher temperatures for the simulations that have the urban LCLU indexes specified in the bottom boundary. This temperature difference occurs, with positive values of up to 2.5°C, mainly downwind of the city core. The spatial pattern of the temperature differences across the SJMA can be explained by the presence of a larger urbanized area in the southwest sector (Figure 3) and persistent northeasterly trade winds during a great portion of the day, which might advect some of the heat stored and released by the downtown area of San Juan on to this region (Figure 11a). All the simulations produced a similar daily cycle for wind patterns, characterized by a strong influence of the northeasterly trade winds.

Precipitation anomalies induced by the interaction of a large urban center with the region’s climatology have been reported in previous studies for continental cities (Bornstein and Lin 2000; Jauregui and Romales 1996; Dixon and Mote 2003) and for coastal locations (Shepherd and Burian 2003; Shepherd 2005). González et al. (González et al. 2005; González et al. 2006) reported observations obtained during the ATLAS Mission where the authors showed that the SJMA was affected by the occurrence of a short and weak precipitation event (∼4 mm of rain) localized in the vicinity of the SJMA on 14 February 2004, and that small amount of precipitation was enough to damp the UHI for that particular day (the reader is directed to the references for these results). In the present study, it was interesting to find that the model was able to capture a small amount of accumulated precipitation, also on the order of ∼4 mm, just southwest of the city in the urban run for 14 February 2004 (Figure 12a). The difference between the total rainfall produced by the urban and the natural simulations show that as much as approximately 30% of this liquid precipitation was not produced in the natural simulation (Figure 12b). Even considering the short duration of the simulations, these results are in agreement for location and order of magnitude, with results obtained in other observational studies. During the Metropolitan Meteorological Experiment (METROMEX), precipitation increases of between 5% and 25% were observed within 50–70 km downwind of various cities (Shepherd 2005).

There are three main factors suggested as possible causes for urban-induced precipitation anomalies: 1) mechanical mixing resulting from increased surface roughness, which will act as a drag on the flow and induce air convergence; 2) the increase of sensible heat flux at the surface, which may lead to a thermally driven convergent mesoscale circulation over large urban areas (Inoue and Kimura 2004); and 3) the release of anthropogenic aerosols into the atmosphere over urbanized areas (Jauregui and Romales 1996). The first two mechanisms are generally favorable for the occurrence of precipitation, because they both enhance upward air motions. However, the third mechanism could act to either enhance or inhibit precipitation, depending on the particle size spectrum of the aerosols being released. All these factors are highly affected by the seasonal changes in insolation and large-scale (synoptic) wind patterns. Because the effects of AP concentration were not included in the simulations, we should consider the combined effect of airflow convergence and the addition of sensible heat when trying to explain the precipitation discrepancies between the two runs.

A more detailed analysis of the mechanical/thermal precipitation-enhancing mechanism using the model results from the urban and natural simulations might provide evidence of the urban influence on the minor precipitation event that occurred during the ATLAS Mission. Here we can see that the addition of sensible heat flux over the SJMA is greatly increased in the urban run for 14 February (Figure 13), whereas the heat fluxes are of the same order over the rural rain forest area for both simulations (not shown). Warm air rises more vigorously and increases the depth of the mixed layer (Figure 14); cumulus clouds may now form as the air parcels become more buoyant and reach their LCL. Some of the clouds may attain appreciable depths such as to produce light precipitation in and around the area. Figure 11a shows that the prevailing northeasterly trade winds could transport these clouds and any accompanying precipitation to the southwest quadrant of the SJMA. In fact, the urban simulation shows a northwest–southeast band of convective cloud development parallel to the coastline with a small enhanced coverage just southwest of the center of the SJMA (Figure 15), which coincides with a region of the model-generated precipitation (Figure 12). This spatial pattern can be explained by the presence of a larger urbanized area in the southwest sector of the SJMA and of the approaching northeasterly trade winds. In the context of numerical modeling, this difference could be attributed to model uncertainties, but its location and occurrence lead us to believe that it is apparently induced by the presence of the urban area and needs to be further studied for longer periods of time.

It is evident that orographic lifting and the northeasterly trade winds exert a significant influence on cloud cover and precipitation over the island of Puerto Rico (Daly et al. 2003; Comarazamy and González 2008). Thus, to better visualize the local effects on the wind pattern across the northeast region of the island, a background wind field was subtracted from the overall wind pattern, and then the wind vector difference at 1500 LST between the urban and natural simulations was calculated. Figure 11b shows the surface horizontal spatial distribution of this computation, whereas Figure 16 shows a vertical cross section of the difference between the urban and natural averaged wind vectors at different times of the day.

Figure 11b shows that the presence of the SJMA tends to accelerate the horizontal wind flow, maybe as a result of the enhanced rising motions in the vertical plane (Figure 16). That is, the enhanced convection induced by the sensible heat increase over the city could be driving a stronger sea-breeze circulation by means of conservation of mass. Figure 11b also suggests that the airflow over the heavily urbanized SJMA may tend to converge, particularly over and downwind of the urban core. We believe that a thermally induced convergence zone developed over the SJMA region and is present in the model results, in spite of the overwhelming influence of the trade winds in modulating the general northeasterly flow characteristics over the city. It is hypothesized that, under a randomly occurring weak large-scale flow regime and similarly clear and stable conditions, a noticeable convergence area over the city might develop. Depending on the stability and moisture profile of the atmosphere, this may lead to modifications in rainfall patterns in and downwind of the SJMA.

Finally, the authors believe that further study of the interaction of the sea–land breeze circulation and a strong tropical coastal UHI, such as the one the SJMA generates, is of the utmost importance to better understand and predict future climate shifts over the city and surrounding areas resulting from increased urbanization. An expanded climatology study focused on this topic is thus recommended for the future.

4. Summary and conclusions

The work presented here is an investigation of the land–atmospheric interaction in a tropical coastal city, including the validation of an atmospheric mesoscale modeling system for use in impact studies. The findings of this research can be summarized as follows:

  • The atmospheric numerical model used (RAMS) was successfully validated to capture the impact of the urban LCLU of San Juan on different atmospheric variables. Daily cycle of air temperatures, vertical profiles of temperatures and wind speeds, and canopy temperature distributions showed good agreement with the observations gathered during the days of the San Juan ATLAS Mission (February 2004).
  • The model land surface configuration was improved using high-resolution airborne remote sensing information to simulate the physical characteristics of the SJMA; to representing its current extension and arrangement, a potential natural vegetation (by interpolating the vegetation surrounding the city to cover the entire area) was also configured. These two configurations were called urban and natural, respectively.
  • The analysis of the two simulated land-use scenarios leads to the conclusion that the urban LCLU have an impact in the general atmospheric dynamics of the northeastern coast of the island of Puerto Rico, especially over and downwind of the SJMA.
  • Model results demonstrate that the influence of the city of San Juan is to produce higher temperatures where the heavily urbanized area is located, as represented by the data retrieved during the ATLAS Mission. This influence could be quantified in air temperatures increases between 2.5° and 3°C.
  • Model results from both the urban and natural runs suggest that the increased sensible heat in the former leads to increased vertical motions and moist convective activity. Bands of clouds and a minimal precipitation event in the urban simulation were traced to occur on areas southwest of the SJMA, likely a result of the weaker northeasterly trade winds advecting the available moisture and increase in energy toward this direction.
  • The spatial pattern of the wind flow characteristics suggests that the ATLAS-represented SJMA strengthens the sea-breeze circulation and may induce convergence over the core of the urbanized region. However, evidence of the latter, which may further enhance rising motions and moist convection, was marginally conclusive.
  • It is worth noting that, between the urban simulation and a preliminary simulation using a homogeneous slab of concrete as the urbanized area, there was a slight reduction of sensible heat flux, which the authors theorize could be related to an increase of soil heat flux in the urban simulation because of having a more realistic urban representation. The thermal response number is proposed as an indicator of heat storage for future studies, because it represents how a surface reacts to different radiative inputs and how it partitions energy.
Future development of the research presented in this paper should include a series of season-long simulations to asses the long-term impacts of environmental changes in the northeast coast of Puerto Rico over the region’s hydrological cycle and its effects on the Caribbean tropical rainforest, located about 15 km east of the SJMA. The season-long simulations should be configured in such a way as to isolate the effects of urbanization and deforestation, industrialization, global warming, and greenhouse gas emissions in the regional climate of Puerto Rico.

Acknowledgments

This research was partially funded by the NASA EPSCoR program of the University of Puerto Rico and by NOAA/CREST Grant NA06OAR4810162. The atmospheric model simulations were performed at the High Performance Computing Facilities in Rio Piedras. Thanks are due to Ana J. Picón and Pieter Van der Mier from UPR for their assistance and help in processing the airborne remote sensing data.

REFERENCES

  • Bornstein, R., and Q. Lin. 2000. Urban heat islands and summertime convective thunderstorm in Atlanta: Three cases studies. Atmos. Environ. 34:507516.

    • Search Google Scholar
    • Export Citation
  • Comarazamy, D. E., and J. E. González. 2008. On the validation of the simulation of early season precipitation in the island of Puerto Rico using a mesoscale atmospheric model. J. Hydrometeor. 9:507520.

    • Search Google Scholar
    • Export Citation
  • Cotton, W. R. Coauthors 2003. RAMS 2001: Current status and future directions. Meteor. Atmos. Phys. 82:529.

  • Daly, C., , E. H. Helmer, , and M. Quiñones. 2003. Mapping the climate of Puerto Rico, Vieques and Culebra. Int. J. Climatol. 23:13591381.

    • Search Google Scholar
    • Export Citation
  • Dickinson, R. E., , A. H. Sellers, , P. J. Kennedy, , and M. F. Wilson. 1986. Biosphere-Atmosphere Transfer Scheme (BATS) for the NCAR Climate Community Model. NCAR Tech. Note NCAR/TN-275+STR. 69 pp.

    • Search Google Scholar
    • Export Citation
  • Dixon, P. G., and T. L. Mote. 2003. Patterns and causes of Atlanta’s urban heat island–initiated precipitation. J. Appl. Meteor. 42:12731284.

    • Search Google Scholar
    • Export Citation
  • González, J. E. Coauthors 2005. Urban heat islands developing in coastal tropical cities. Eos, Trans. Amer. Geophys. Union 86.doi:10.1029/2005EO420001.

    • Search Google Scholar
    • Export Citation
  • González, J. E., , J. C. Luvall, , D. Rickman, , D. E. Comarazamy, , and A. J. Picón. 2006. Urban heat island identification and climatological analysis in a coastal, tropical city: San Juan, Puerto Rico. Urban Remote Sensing, Q. Weng and D. Quattrochi, Eds., CRC Press, 223–252.

    • Search Google Scholar
    • Export Citation
  • Inoue, T., and F. Kimura. 2004. Urban effects on low-level clouds around the Tokyo metropolitan area on clear summer days. Geophys. Res. Lett. 31:L05103. doi:10.1029/2003GL018908.

    • Search Google Scholar
    • Export Citation
  • Jauregui, E. 1997. Heat island development in Mexico City. Atmos. Environ. 31:38213831.

  • Jauregui, E., and E. Romales. 1996. Urban effects of convective precipitation in Mexico City. Atmos. Environ. 30:33833389.

  • Kalnay, E. Coauthors 1996. The NCEP/NCAR 40-Year Reanalysis Project. Bull. Amer. Meteor. Soc. 77:437471.

  • Landsberg, H. E. 1981. The Urban Climate. Academic Press, 275 pp.

  • Lawton, R. O., , U. S. Nair, , R. A. Pielke Sr., , and R. M. Welch. 2001. Climatic impact of tropical lowland deforestation on nearby montane cloud forests. Science 294:584587.

    • Search Google Scholar
    • Export Citation
  • Luvall, J. C., and H. R. Holbo. 1989. Measurement of short-term thermal responses of coniferous forest canopies using thermal scanner data. Remote Sens. Environ. 27:110.

    • Search Google Scholar
    • Export Citation
  • Luvall, J. C., , D. Lieberman, , M. Lieberman, , G. S. Hartshorn, , and P. Peralta. 1990. Estimation of tropical forest canopy temperatures, thermal response numbers, and evapotranspiration using an aircraft-based thermal sensor. Photogramm. Eng. Remote Sens. 56:13931401.

    • Search Google Scholar
    • Export Citation
  • Meyers, M. P., , R. L. Walko, , J. Y. Harrington, , and W. R. Cotton. 1997. New RAMS cloud microphysics parameterization. Part II: The two-moment scheme. Atmos. Res. 45:339.

    • Search Google Scholar
    • Export Citation
  • Nair, U. S., , R. O. Lawton, , R. M. Welch, , and R. A. Pielke Sr.. 2003. Impact of land use on Costa Rican tropical montane cloud forests: Sensitivity of cumulus cloud field characteristics to lowland deforestation. J. Geophys. Res. 108:4206. doi:10.1029/2001JD001135.

    • Search Google Scholar
    • Export Citation
  • Noto, K. 1996. Dependence of heat island phenomena on stable stratification and heat quantity in calm environment. Atmos. Environ. 30:475485.

    • Search Google Scholar
    • Export Citation
  • Oke, T. R. 1987. Boundary Layer Climates. Cambridge University Press, 450 pp.

  • Pielke, R. A. Coauthors 1992. A comprehensive meteorological modeling system—RAMS. Meteor. Atmos. Phys. 49:6991.

  • Poreh, M. 1996. Investigation of heat islands using small scale models. Atmos. Environ. 30:467474.

  • Ray, D. K., , U. S. Nair, , R. O. Lawton, , R. M. Welch, , and R. A. Pielke Sr.. 2006. Impact of land use on Costa Rican tropical montane cloud forests: Sensitivity of orographic cloud formation to deforestation in the plains. J. Geophys. Res. 111:D02108. doi:10.1029/2005JD006096.

    • Search Google Scholar
    • Export Citation
  • Rickman, D. L., , J. C. Luvall, , and S. Schiller. 2000. An algorithm to atmospherically correct visible and thermal airborne imagery. Workshop on Multi/Hyperspectral Sensors, Measurements, Modeling, and Simulation, Huntsville, Alabama, U.S. Army Aviation & Missile Command, SR-RD-WS-01-01.

    • Search Google Scholar
    • Export Citation
  • Rozoff, C. M., , W. R. Cotton, , and J. O. Odegoke. 2003. Simulation of St. Louis, Missouri, land use impacts on thunderstorms. J. Appl. Meteor. 42:716738.

    • Search Google Scholar
    • Export Citation
  • Saleeby, M. S., and W. R. Cotton. 2004. A large-droplet mode and prognostic number concentration of cloud droplets in the Colorado State University Regional Atmospheric Modeling System (RAMS). Part I: Module descriptions and supercell test simulations. J. Appl. Meteor. 43:182195.

    • Search Google Scholar
    • Export Citation
  • Shepherd, J. M. 2005. A review of current investigations of urban-induced rainfall and recommendations for the future. Earth Interactions 9. [Available online at http://EarthInteractions.org].

    • Search Google Scholar
    • Export Citation
  • Shepherd, J. M., and S. J. Burian. 2003. Detection of urban-induced rainfall anomalies in a major coastal city. Earth Interactions 7. [Available online at http://EarthInteractions.org].

    • Search Google Scholar
    • Export Citation
  • Taylor, M. A., , D. B. Enfield, , and A. A. Chen. 2002. Influence of the tropical Atlantic versus the tropical Pacific on Caribbean rainfall. J. Geophys. Res. 107:3127. doi:10.1029/2001JC001097.

    • Search Google Scholar
    • Export Citation
  • Tso, C. P. 1996. A survey of urban heat island studies in two tropical cities. Atmos. Environ. 30:507519.

  • Van der Molen, M. K. 2002. Meteorological impacts of land use change in the maritime tropic. Ph.D. thesis, Vrije Universititeit Amsterdam, 262 pp.

  • Van der Molen, M. K., , A. J. Dolman, , M. J. Waterloo, , and L. A. Bruijnzeel. 2006. Climate is affected more by maritime then by continental land use change: A multiple scale analysis. Global Planet. Change 54:128149.

    • Search Google Scholar
    • Export Citation
  • Velazquez-Lozada, A., , J. E. Gonzalez, , and A. Winter. 2006. Urban heat island effect analysis in San Juan, Puerto Rico. Atmos. Environ. 40:17311741.

    • Search Google Scholar
    • Export Citation
Figure 1.
Figure 1.

Temperature profile observed in the SJMA, based on data collected during the ATLAS Mission (February 2004) from surface weather stations and temperature sensors located along the east–west solid line in the bottom panel of Figure 3 and constructed with the average daytime temperatures.

Citation: Earth Interactions 14, 16; 10.1175/2010EI309.1

Figure 2.
Figure 2.

Model grids used in the numerical simulations. The topography contours have intervals of 150 m.

Citation: Earth Interactions 14, 16; 10.1175/2010EI309.1

Figure 3.
Figure 3.

Specification of the surface characteristics used in the two atmospheric model runs used for the study for the (top) natural and (bottom) urban scenarios. The reader is referred to Table 1 for a description of each class number and its corresponding biophysical parameters. The solid line in the urban scenario represents the transect line where the stations and sensors used to construct Figure 1 are located.

Citation: Earth Interactions 14, 16; 10.1175/2010EI309.1

Figure 4.
Figure 4.

RAMS/ATLAS overall data processing flow to produce atmospheric, radiometric, and geometrically corrected data and link to update the mesoscale model surface characteristics (adapted from Rickman et al. 2000).

Citation: Earth Interactions 14, 16; 10.1175/2010EI309.1

Figure 5.
Figure 5.

ATLAS sensor visible bands 1–3: this information was used to identify the extension and morphology of the SJMA.

Citation: Earth Interactions 14, 16; 10.1175/2010EI309.1

Figure 6.
Figure 6.

ATLAS-derived albedo dataset used to configure the surface characteristics of RAMS.

Citation: Earth Interactions 14, 16; 10.1175/2010EI309.1

Figure 7.
Figure 7.

Comparison of canopy temperatures between those simulated by RAMS and those observed by the ATLAS sensor when it was flown over the SJMA.

Citation: Earth Interactions 14, 16; 10.1175/2010EI309.1

Figure 8.
Figure 8.

Comparison of the air temperatures between the regional model results and the stations and sensors deployed in the SJMA.

Citation: Earth Interactions 14, 16; 10.1175/2010EI309.1

Figure 9.
Figure 9.

Vertical profiles of temperature (°C) and wind speed (m s−1) for model results [solid line is SJMA ATLAS; dashed line is SJMA homogeneous slab] and the San Juan NWS office balloon data (open squares) for (a) 1200 UTC 11 Feb, (b) 0000 UTC 14 Feb, and (c) 1200 UTC 16 Feb 2004.

Citation: Earth Interactions 14, 16; 10.1175/2010EI309.1

Figure 10.
Figure 10.

Spatial distribution of the air temperature difference (°C) at 2 m AGL between the urban and natural scenarios simulated for the analysis.

Citation: Earth Interactions 14, 16; 10.1175/2010EI309.1

Figure 11.
Figure 11.

(left) Wind field averaged at 1500 LST during the complete period of simulation of the present run. (right) Model wind field difference between the urban and natural runs averaged at 1500 LST.

Citation: Earth Interactions 14, 16; 10.1175/2010EI309.1

Figure 12.
Figure 12.

(left) Total accumulated precipitation for the urban run during the ATLAS period (in mm and limited to the 9-mm contour because the El Yunque total, >50 mm, masked precipitation in other regions) and (right) precipitation difference (in percentage increase) between the urban and natural runs.

Citation: Earth Interactions 14, 16; 10.1175/2010EI309.1

Figure 13.
Figure 13.

Increased sensible heat flux over the urban area for the urban and natural scenario simulations.

Citation: Earth Interactions 14, 16; 10.1175/2010EI309.1

Figure 14.
Figure 14.

Simulated vertical motion fields (10−1 m s−1) during 14 Feb 2004 at the specified times for the (top) urban and (bottom) natural runs. Early morning hour panels were omitted because of the suppressed vertical motions that are typical for calm and stable conditions at this time of the day.

Citation: Earth Interactions 14, 16; 10.1175/2010EI309.1

Figure 15.
Figure 15.

Fractional cloud coverage predicted by RAMS for the urban run at selected times on 14 Feb 2004.

Citation: Earth Interactions 14, 16; 10.1175/2010EI309.1

Figure 16.
Figure 16.

North–south vertical cross sections of wind vectors difference between the urban and natural runs for the average (top left)–(top right) 1200, 1400, and 1600 LST and (bottom left)–(bottom right) 1800, 2000, and 2200 LST. The thick horizontal line at the bottom of each panel depicts the presence of land, and the two thick vertical lines represent the two reference topographic peaks of ∼700 and 200 m.

Citation: Earth Interactions 14, 16; 10.1175/2010EI309.1

Table 1.

LEAF-2 biophysical parameters by land-use class number.

Table 1.
Save