Derivation of Spatially Distributed Thermal Comfort Levels in Jordan as Investigated From Remote Sensing, GIS Tools and Computational Methods

Thermal comfort is usually calculated using discrete point measurements. This procedure is not 4 suitable to study thermal comfort for inhabited areas with rugged terrains where climate gradient is 5 high. The wide availability of remote sensing data and GIS tools have revolutionized data management, 6 processing and visualization. The present paper implemented digital elevation data, GIS tools and a 7 computational algorithm to generate spatially continuous maps of climatological elements which were 8 employed to derive thermal comfort levels across Jordan. Results show detailed information of the 9 spatial distribution of the degree of thermal comfort in winter and summer across the country which 10 cannot be resolved using discrete point measurements. It is shown that the mountainous areas in the 11 country, where most urban centers are situated, experience “ slightly warm ” to “ warm ” indoor apparent 12 temperatures in summer. The Jordan Valley and the desert experience high indoor apparent 13 temperatures in summer. Cold conditions prevail over most parts of the country, with the heating 14 degree days ranging from 2100 in the southern mountains to values close to zero near the Dead Sea 15 area. The presented procedure demonstrated that the very low levels of ambient vapor pressure is an 16 important atmospheric forcing contributing to the widespread cold conditions prevailing over the desert 17 areas in winter. The efficiency of direct evaporative cooling systems to achieve thermal comfort in the 18 various parts of the country is investigated. The procedure presented can be used over regional scales 19 with different levels of spatial resolutions for a wide range of climatological studies.

The present procedure is applied to the country of Jordan. Jordan extends from 29.19 O Table 1 shows the coordinates, elevation, 84 average monthly air temperature and relative humidity of the stations used in this investigation. To 85 improve the integrity of the model and provide a better representation of the study area, five additional 86 stations were synthetically generated. This was done primarily along the Dead Sea and in Wadi Araba station were derived based on field measurements and a few other criteria including its elevation, the general climate pattern of the nearest measuring stations, wind regime and energy balance considerations (Oroud, 1997).
Interpolation by GIS tools is purely spatial, and thus climate gradient induced by elevation differences across a landscape is usually masked out. The problem of GIS interpolation can become severe if 94 weather measuring stations are sparsely distributed over very rugged terrains (e.g., Jeffrey et al., 2001; produce adequate results, and large deviations may result. Under such conditions, a better procedure to 97 generate accurate data between measuring sites would be to build a numerical procedure which takes 98 into account the neighboring measuring sites along with vertical and horizontal gradient of the intended 99 climate elements across the landscape. The computational procedure provides an explicit algorithm to 100 calculate meteorological elements at each grid point. This procedure produces a dense point network 101 which enables GIS to produce accurate spatially distributed raster database.

102
Two steps are needed to generate climatological data over a spatial grid. The one is obtaining the distance between the grid and measuring sites, and the second one is calculating the climatological 104 element using the closest n stations after taking into account the elevation difference between each 105 measuring site and the spatial grid.

106
The Euclidean distance between grid j and a measuring site is computed as, where δi,k is the Euclidean distance (m) between station i and the intended grid cell, and the x and y are 110 the easting and northing coordinates (m).

111
The temperature of each grid (T(xk,yk,zk) was calculated explicitly using the inverse distance weighting 112 procedure (IDW) using the following equation (Emmendorfer and Dimuro, 2020), where xk is the easting (m), yk is the northing (m) and zk is the elevation of the grid; Ti is the the elevation of the grid, and δ(i,k) is the Euclidean distance between station i and the intended cell. In 117 this procedure, the closest three stations were chosen to carry out the interpolation of temperature for 118 each grid (the j parameter in Eqn. 2).

119
The actual vapor pressure is computed from the interpolated relative humidity and saturation water 121 vapor pressure at each grid, where RH is the relative humidity (%), ea and es are the actual and saturated vapor pressure (kPa),

124
respectively. The saturation vapor pressure at a grid is calculated using the following form (Oroud,

127
Where T is the grid temperature ( o C).

128
The wet bulb temperature which is an important indicator of the evaporating power of air is usually 129 used to assess the efficiency of using evaporative cooling systems. The wet bulb temperature is      shows the spatial distribution of average monthly air temperature during January and July, respectively.

184
Wide temperature differences are observed across the study area during these two months.

192
The summer months, as represented by July, display a different spatial temperature distribution pattern  The spatial distribution of indoor apparent temperature is presented in Figure 5 for January and July as

227
Two additional atmospheric factors enhance nighttime cooling-the first one is the usual absence of 228 cloud cover, and the second factor is the low moisture content in the troposphere. Figure 6 shows the

295
The present paper implemented a combination of tools to generate high spatial resolution 296 meteorological data across a regional scale. The generated data were extended to derive the spatial   404 Figure 4. The spatial distribution of average monthly air temperature in July and January.
405 Figure 5. The spatial distribution of indoor apparent temperature in July and January.
406 Figure 6. The spatial distribution of ambient vapor pressure across Jordan in January.    Topography of study area along with the measuring climatological stations. Flowchart of steps used to carry out the investigation.

Figure 4
The spatial distribution of average monthly air temperature in July and January.

Figure 5
The spatial distribution of indoor apparent temperature in July and January.

Figure 6
The spatial distribution of ambient vapor pressure across Jordan in January.

Figure 7
A scatterplot of the relationship between indoor thermal comfort and elevation.

Figure 8
The average monthly indoor apparent temperature for twenty-two locations.

Figure 9
Spatial distribution of the wet bulb temperature during July.