Notice: Wiley Online Library will be unavailable on Saturday 27th February from 09:00-14:00 GMT / 04:00-09:00 EST / 17:00-22:00 SGT for essential maintenance. Apologies for the inconvenience.
Corresponding author: A. H. Maghrabi, National Centre for Mathematics and Physics, King Abdulaziz City for Science and Technology, PO Box 6086, Riyadh 11442, Saudi Arabia. (email@example.com)
 Sky temperatures that were estimated from a single-channel IR detector over Riyadh, Saudi Arabia, were analyzed from June 2008 to May 2011. The data were divided into three main categories: clear sky, cloudy sky, and dusty conditions. The observation and the research results were as follows. During periods of clear-sky conditions, it was found that the sky temperatures depend mainly on the atmospheric water content, the screen level temperature, and the suspended aerosol particles in the atmosphere. Under cloudy conditions, the sky temperature ranges between −37°C and 5°C. The mean sky temperatures in this case are higher than those of the clear-sky conditions by approximately 11°C to 18°C. The radiative properties of cloudy skies depend on the cloud characteristics and the intervening atmosphere between the ground and the cloud base. The sky temperature during dusty conditions ranged between −20°C and 8.5°C. The study showed that dusty conditions increase the atmospheric temperatures by approximately 17°C to 31°C. The sky temperatures during dusty periods are affected by several factors, such as the air mass properties, which bring the dust, and the dust particle characteristics, such as size, shape, and chemical composition, which are initially determined by the sources from which the dust originated. Theoretical simulations using MODTRAN software were used to investigate the atmospheric thermal radiation spectral distributions in the three categories. The results show that the major changes occurred within the atmospheric window (8–14μm).
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
 Thermal radiation is due to the presence of gases in the atmosphere that absorb and emit energy in the infrared (IR) wavelength. This absorption/emission in the atmosphere is mainly due to water vapor (below 7.6 μm, between 13 μm and 16 μm, and beyond 22 μm), carbon dioxide (from 14 μm to 16 μm) and, to a lesser extent, ozone (between 9 and 10 μm). Trace gases, such as methane, nitrous oxides, and carbon monoxide, also absorb and emit in the IR wavelengths. The remaining unabsorbed portion of the Earth's radiation escapes into outer space through the AW, which lies between 8 μm and 14 μm (Figure 1) [Goddy and Yung, 1989].
 Other factors, such as suspended aerosol particles, are expected to affect atmospheric thermal radiation. Their effect varies in importance according to several parameters, such as the wavelength and the atmospheric conditions. The impact of suspended aerosol particles on atmospheric thermal radiation mainly occurs within the AW, which causes an increase in the total atmospheric thermal radiation.
 The purpose of this paper is to study the behavior of atmospheric thermal radiation (represented here by the sky temperature, Tsky) under clear and cloudy sky conditions and to compare it with those of dusty conditions.
 The observational study area, the data set, and the methodology are summarized in section 2. The results are presented and discussed in section 3, and the conclusions are given in section 4.
2. Instrumentation and Methodology
 The data used in this study consist of sky temperature measurements, surface meteorological data, and upper air temperature and humidity profiles from radiosondes.
 The sky temperature measurements were obtained from a single-pixel IR detector with a field of view of 3° and a spectral range from 5.5 to 50μm. The principles of this detector, its constructions, its calibration procedures, and its performances have been described previously in Maghrabi et al. , Maghrabi and Clay , and Clay et al. .
 The detector is based on commercial IR thermopile sensors (EG&G Heimann TPS 534). The thermopile has a wide spectral range from 1 μm to approximately 50 μm. The sensor elements and the required electronics are hermetically sealed into a housing, and an additional optical filter is used to better define the optical passband. We use an optical filter with a 5.5 μm spectral short-wave cutoff.
 The electronics of the detector provide three outputs: the uncompensated temperature, the ambient temperature of the detector, and the compensated temperature. The three outputs are related to each other by the following relation: Tskyuncompensated = Tskycompensated − Tambient, where Tskyuncompensated is the relative sky temperature that is, the temperature difference between the target compensated temperature (Tskycompensated) and the detector reference temperature (Tambient). In this paper, all the presented results and the given discussion, we use the relative sky temperature as a measure of the sky temperature and will be referred to as Tsky.
 A blackbody (BB) of known temperature was used to calibrate the detector's outputs. The experimental error in the temperature measurements from the detector was found to be approximately 1°C. [Riordan et al., 2005; Maghrabi et al., 2009].
 The sky temperatures were acquired every 15 min using an XR5-8-A-SE data logger manufactured by Pace Scientific. The logger has internal sensors to measure humidity and air temperature. The accuracy of the logger's sensor measurements is ±2% humidity and ±0.15°C at a temperature of 25°C.
 The detector was placed on the roof of the King Abdulaziz City for Science and Technology (KACST) main building in Riyadh (24°43′N, 46°40′E, 764 m above sea level), Saudi Arabia for the period between 2008 and 2011. However, due to the difficulty of maintaining yearlong measurements, there have been cases of missing data for some time points during the period of measurements. Moreover, the detector was taken down for maintenance and calibration procedures during some periods. The available data cover the periods between April and June 2008, from November 2008 to April 2009, from September 2009 to October 2010, and from February to May 2011. The data obtained during these periods of time cover almost all four seasons.
 Hourly observations of meteorological variables, such as air temperature, relative humidity, visibility, and sky conditions, which were recorded at Riyadh Airport (10 km from the KACST site), were provided by the Presidency of Meteorological and Environment (PME). The upper air temperature and humidity profiles from the radiosonde, which was launched twice a day at Riyadh Airport, were also available.
 The sky condition, as recorded by PME observers, describes the sky at the time of observations. The recorded conditions include cloudy periods (e.g., overcast, partly cloudy, and scattered clouds), clear-sky times, dust storm events (e.g., sever storm, dust storm, and blowing dust), rain, thunder storm, drizzle, and haze.
 To study the impact of dusty conditions on sky temperature and to compare it with other atmospheric conditions, the data set was divided into three main groups based on the sky condition records provided by PME: clear sky, cloudy sky, and dusty conditions.
 Clear-sky times can be defined as the time in which no clouds and no dusty event were observable. During these times, the visibility is expected to be greater than 10 km.
 Cloudy-sky conditions were categorized into three subgroups according to the information provided by the PME: overcast, partly overcast and scattered cloud conditions. Overcast skies are usually characterized by cloud coverage of approximately 7–8 octa, partly cloudy skies have between 4 and 7 octa, and skies with patches of clouds (scattered clouds) have cloud coverage of 2–4 octa.
 Dusty conditions were also divided into three categories: blowing dust (BD), dust storm (DS), and severe dust storm (SS). Kutiel and Furman  have defined these conditions according to the reduction of the visibility as follows. Blowing dust represents raised dust or sand at the time of observation and usually reduces visibility between 1 to 6 km. A dust storm occurs when strong winds lift large quantities of dust particles, which reduces the visibility between 200 m and 1000 m. A severe dust storm occurs when very strong winds lift large quantities of dust particles, which reduces the visibility to less than 200 m. As an example, Figure 2 shows the atmospheric situation experienced in Riyadh during each of the three types of dusty conditions.
 To avoid any overlap between dusty and cloudy conditions, the sky has to be clear for at least 2 h before the conditions are considered to be dusty or cloudy. If two events occurred on the same day (dust and clouds), they have to be separated by at least 4 h of clear-sky time for both conditions to be recorded. If the separated time between the two events was less than 4 h, the first condition will be considered, and the second will be ignored. For instance, if the sky was reported to be clear between 04:00 and 09:00, cloudy between 09:30 and 12:00, dusty between 13:30 and 15:00, clear between 16:00 and 21:00 and cloudy again between 21:00 and 23:00, then, for this particular day, we consider the sky as cloudy for the times between 9:30 and 12:00 and between 21:00 and 23:00., and the dusty periods will not be considered. This method was applied for the other atmospheric conditions, and it allows us to reduce the effects of one type of the sky on another.
Table 1 shows the number of measured hours for each sky condition type. The mean, maximum, and minimum sky temperatures values and the visibilities for dusty sky are compared with those for clear and cloudy skies. Figure 3 shows the frequency distribution of the sky temperatures for each sky condition.
Table 1. Summary of the Mean, Maximum, and Minimum Sky Temperatures and the Observed Visibility for the Seven Atmospheric Conditions Considered During the Number of Hours in the Study Period
Number of Hours
Mean Temperature (°C)
Minimum Temperature (°C)
Maximum Temperature (°C)
Over Cast (OC)
Blowing Dust (BD)
Dust Storm (DS)
Sever Storm (SS)
 Theoretical simulations were conducted using MODTRAN software version3 Revision1 (Mod4v3r1) [Berk et al., 1989]. The MODTRAN program is a rigorous radiation transfer algorithm that is used to calculate, at moderate resolution, the spectral absorption, transmission, emission, and scattering characteristics of the atmosphere for wavelengths extending from thermal infrared through the visible and into the ultraviolet. MODTRAN was developed by the U.S. Air Force Research Laboratory (AFRL) program in the early 1970s, and it has been used for various applications and studies [Berk et al., 1989].
 The main objectives of using MODTRAN in the current study are to investigate the spectral distribution of atmospheric thermal radiation under different atmospheric conditions and to explain and justify the experimental findings.
 The steps in the simulation methodology and the approach for the calculations are as follows.
 1. Choose the desired atmospheric profile. For this purpose, MODTRAN was run with its standard atmospheric profiles, mainly midlatitude summer (MLS), and real radiosonde atmospheric profiles for Riyadh.
 2. Assign the necessary input parameters for a certain investigation. For example, to investigate the effect of different types of cloud on atmospheric thermal radiation, the aerosol types were set at a fixed value, and the cloud type was changed in each run. Similarly, at a fixed amount of water content, different aerosol types and their effect on atmospheric thermal radiation were examined.
 3. Execute MODTRAN.
 4. Process the outputs. MODTRAN calculates the atmospheric thermal radiation in the wavelength interval between 4 and 100 μm in 0.1 μm steps and outputs the calculated spectral radiance in (W str−1 cm−2μm−1). We have developed a simple algorithm to interface with the MODTRAN output files and to calculate the sky temperatures. The sky temperatures were obtained by integrating the spectral radiance between the wavelengths of 5.5 μm to 50 μm (the spectral range of our detector). Then, assuming the atmosphere as a uniform blackbody with a uniform sky temperature and using the Stefan-Boltzmann law, the integrated radiance is converted into temperature, which is conventional for the purposes of investigation conducted in this study. All MODTRAN obtained temperatures are relative to the air (screen level) temperatures.
3. Results and Discussion
3.1. One Clear-Sky Temperature
Figure 3a shows the frequency distribution of the sky temperatures for the total 6511 h of measurements for clear skies. The sky temperature ranges between −48°C and −2°C, with a mean value of −28°C. This range of sky temperature variation is attributed to changing atmospheric conditions, e.g., the amount of water vapor content and/or the air temperature [Dilley and O'Brien, 1998; Ruckstuhl et al., 2007; Dupont et al., 2008; Prata, 1996]. The presence of small aerosols particles from sources other than dust storms may also have some effect on atmospheric radiation, especially when the size of the particles is comparable to the IR wavelengths.
Figure 4 shows the relationship between the sky temperature and both the water content (represented by the vapor pressure; Figure 4a) and the air temperature (Figure 4b) for the period between November 2008 to April 2009. Although there was some scatter in the data, the sky temperatures are well correlated with both parameters. As shown in Figure 4, we found that the sky temperature increases by 2.2°C and 1.2°C when the vapor pressure increases by 1 mbar and the air temperature increases by 1°C.
 The relationship between the air and the sky temperatures indicates that the temperature in the part of the atmosphere where the radiation originates is correlated, via some combination of radiation, turbulence and advective process, with the air temperature. The dependence of the sky temperature on the second parameter (water vapor content) is generally needed, since atmospheric water vapor strongly absorbs the infrared radiation [Brunt, 1932; Swinbank, 1963; Brutsaert, 1975; Idso, 1981; Ruckstuhl et al., 2007; Maghrabi and Clay, 2011].
Figure 5shows the atmospheric thermal radiation spectral distribution obtained by MODTRAN for various amounts of water content, represented here by precipitable water vapor (PWV). Here, we executed MODTRAN with its standard MLS atmosphere for different amounts of PWV. There is an evident increase in the area under the curves of the atmospheric thermal radiation as the amount of water is increased. As the amount of PWV increases, the emission from the water vapor first approaches the BB emission in the spectral region on each side of the AW. A further increase in PWV is accompanied by an increase in emission inside the AW toward the BB values at the screen-level temperature. In this example, the change in PWV values between 0.05 and 4 cm (vapor pressure between 0.003 and 24 mbar) increases the atmospheric radiation by approximately 85%.
Figure 6shows the relationship between the measured temperature and the MODTRAN-simulated temperature. Here, the pressure, the temperature, and the relative humidity for 249 profiles from radiosonde operated at Riyadh Airport were used as a user-defined atmospheric model input into MODTRAN. The profiles of the other atmospheric constituents (e.g., CH4, CO, O2, and NO) were set to those of the MODTRAN standard MLS model. The straight line in Figure 6 is the 1:1 relation between the two variables. The data lie very close to the 1:1 line with the mean bias error (MBE) and the root mean square error (RMSE) between the measured and the simulated temperatures of 0.8°C and 3°C, respectively. The maximum MBE values were 5.5°C and −6.1°C, which were only reported on some rare occasions. There are several causes for such large deviations. For example, some measurements were made during the daytime; at these times, the solar radiation heats up the IR transmitting optics of the detector, leading to some bias in the measurements [Alados-Arboledos et al., 1988; Philipona et al., 2001; Philipona et al., 1995]. For our detector, this bias is estimated to be approximately 2–3°C [Maghrabi and Clay, 2011].
 To investigate the effect of suspended aerosol particles on the atmospheric thermal radiation spectrum, MODTRAN was executed with its MLS atmosphere for different aerosol types, as shown in Figure 7. The different aerosol types are as follows: no aerosol, rural with visibility = 23 km, tropospheric with visibility = 50 km, urban with visibility = 5 km, maritime with visibility = 23 km, and fog with visibility = 0.5 km. The foggy condition represents the maximum thermal emission, whereas the no-aerosol condition represents the maximum clarity. In foggy conditions, the atmosphere emits as a full BB due to the presence of a huge amount of water. However, for the other aerosol types, there are no major changes outside the atmospheric window. Within the AW, the atmospheric emission increases slightly as the type of aerosol changes to its highest value for urban aerosol atmosphere, which has a visibility of 5 km. This rate of increase is very small in comparison with the effect of higher water values, clouds, or high turbidity values, as we shall see later. The integrated sky temperature over the considered wavelengths (5.5 to 50μm) changes by a maximum of 3°C as we change from the tropospheric aerosol type with visibility = 50 km to the urban aerosol type with visibility = 5 km. We conclude that under normal conditions, the aerosol particles in the atmosphere have little influence on the atmospheric radiation. However, in extreme conditions, we shall see that the situation is different.
3.2. Cloudy Skies
 The frequency distributions for the three cloudy conditions are presented in Figures 3b–3d. The sky temperature of the overcast sky condition (Figure 3b) ranges between −26.5°C to approximately 5.3°C with an average temperature of −10.2°C. The radiative properties of the sky in cloudy conditions depend on a set of parameters related to cloud type and structure, including height, water content, and droplet size distribution. Moreover, the properties of the intervening atmosphere between the ground and the cloud base may also have some effect on the sky temperature variations. Clouds have a strong effect on atmospheric thermal radiation, and this influence decreases in importance with altitude because higher clouds are usually colder than lower clouds [Sugita and Brutsaert, 1993; Riordan et al., 2005].
 In partly overcast skies and scattered cloudy conditions, clouds and clear sky present at the same time (Figures 3c and 3d). The sky temperatures in this case are affected by the atmospheric conditions of the clear skies, by the amount of cloud cover in the sky, and by the cloud properties. The mean value of the sky temperatures for partly and scattered cloudy skies are intermediate between those of clear sky and overcast sky. The partly overcast skies are warmer than the scattered clouds condition and the clear-sky condition by approximately 3°C and 14°C, respectively, and colder than the overcast skies by 4°C. Generally speaking, the sky temperatures in cloudy conditions are shifted toward warmer temperatures as the amount of cloud cover increases.
Figure 8shows the hourly variations of the sky temperature taken between 26 and 30 April 2009 under different atmospheric conditions. The sky experienced both clear and cloudy times over this period. The cloudy condition varies between totally overcast to a sky with little patches of clouds. The sky stayed clear on the first day, from the afternoon of 26 April until 13:00 the next day. During this time, the sky temperature remained between −27 and −35°C. A sharp increase to approximately −14°C in the morning of 27 April was due to the presence of some patches of low-level clouds, which cleared after only an hour. After this point (from midday of 27 April), there was a buildup of clouds observed. The sky temperature starts to increase from its background level. During this cloudy period (13:00 on 27 April to 01:00 on 28 April), there were overcast periods where the sky temperature remained between −10 to −15°C. From 10:00 on 28 April, the sky remained clear for more than 12 h, during which the sky temperatures were between −35°C and −42°C. After that and for the rest of the period, the sky experienced a mix of cloudy periods with different amounts of cloud coverage and some periods of clear sky.
 The effect of clouds at different altitudes on the atmospheric thermal radiation spectral distribution was investigated using MODTRAN with the standard MLS atmosphere. Figure 9 shows the atmospheric thermal radiation spectral distribution in the presence of three types of clouds: stratus (low cloud surface, 2,000 m), cirrus (high clouds, 5,000–13,000 m), and altostratus (middle clouds, 2,000–7,000 m). Due to their higher altitude, the spectrum of the cirrus clouds is similar to that of clear skies (e.g., Figure 1). Meanwhile, the basic effect of the low- and middle-level clouds is to close off the AW to atmospheric thermal radiation and to emit as a BB with a temperature close to that of the screen level. The emission degrees of these two types of clouds differ from each other and depend on the physical properties and the height of the cloud.
 To examine the theoretical results of the sky temperature under cloudy conditions, cloud information such as the cloud height and the cloud type are needed in MODTRAN. Unfortunately, such data were not available from the available meteorological records. To overcome this difficulty, we have estimated the cloud height from the radiosonde profiles of temperature, relative humidity, and dew point depression using the methods described by [Riordan et al., 2005; Arabey, 1975; Chernykh and Eskridge, 1996]. The clouds heights for the 59 cases of overcast conditions were determined to be between 200 m and approximately 9000 m. Once the cloud height was obtained, it was possible to choose the cloud type from the MODTRAN options. We executed MODTRAN with the derived cloud height and the corresponding MODTRAN cloud type.
Figure 10 shows the measured sky temperature plotted against the simulated temperatures obtained from MODTRAN for the 59 totally overcast skies. There is a great agreement between the two temperatures; hence, the data lie close to the 1:1 line. The MBE and RMSE between the two temperatures were 0.5°C and 1.5°C, respectively.
3.3. Dust Storms
3.3.1. Dust Storm Distribution and Frequency
Figure 11 shows the number of hourly sky temperatures for the three categories of dust storms during the study period. It is evident that Riyadh had a dust storm every month in these two years.
 The strength and the occurrence rate for each dust event from each category differ from month to month. Although the four types of storm occur in different months, all of them peaked in the premonsoon season. The number of dusty events for all categories increased from January toward their maximum in March.
3.3.2. Dust Storm Affecting Sky Temperature
 The statistical parameters of the dusty events are presented in Table 1, and their distributions are illustrated in the last 3 histograms of Figure 3. For a total of 1160 h of BD, the sky temperature ranges between −21°C and 0°C with an average value of −11.5°C. For the dust storm with 109 observed hours, the temperatures lie between −18°C and 4°C, with an average value of about −6°C.
 The wide ranges of sky temperatures in dusty conditions can be due to several factors. For instance, the properties of the air masses that brought the dust storms to the region can lead to such variations [ Alharbi, 2009; Badarinath et al., 2007]. Moreover, the impact of dust aerosols in the atmosphere on the sky temperature depends mainly on the particle characteristics such as size, shape, chemical composition and mineralogy [Dayan et al., 1991]. While these characteristics can change during dust transport [Badarinath et al., 2007; Kambezidis and Kaskaoutis, 2008; Kutiel and Furman, 2003], they are initially determined by their source regions. This finding needs more investigation in the future, but it is beyond the scope of this study to investigate the correlations between dust aerosol sizes and air mass sources and the variations of the sky temperatures.
 It is also interesting to note that the average sky temperature during BD is comparable to that in overcast conditions, and it is higher than the average temperature of partly overcast skies and of skies with scattered clouds by 3°C and 6°C, respectively.
 In the case of SS, the situation is different from that under the other dusty conditions. The sky temperature confined between −3°C to 8.5°C with an average value of approximately 3°C, which is approximately 31°C higher than the mean temperature of clear skies and approximately 13°C higher than the mean sky temperature of totally overcast skies. This result implies that such dusty events tend to warm up the atmosphere and in some situations, such as those during SS or DS, the effect of these events become more than that of the clouds. Consequently, airborne particles from dust storms alter the local climatic conditions by modifying the energy budget through their behavior and causing heating to the atmosphere.
Figure 12shows the variations of the sky temperature during two dusty periods; the first one is for a DS event occurring on 26 March 2010, and the second is for an SS event on 19 March 2010. In both examples, the sky temperature increased as a consequence of the dust events, and the sky temperature increase was more than 15°C. However, the two events differ from each other by their response to the event. In the latter case, the dust plume was much stronger than that in the former case, and the sky temperature changes dramatically. Additionally, the amount of aerosol particles brought by the storm and the duration of the event may differentiate the two events from each other. Moreover, the recovery time of the event depends on the severity of the storm and its duration. In some cases, short-lasting SS events have more effect than long-lasting DS events [Maghrabi et al., 2011].
 The effect of aerosol and dust particles on the atmospheric thermal radiation spectrum was examined for 14 dusty events that occurred in Riyadh using MODTRAN. The upper air data from the radiosonde for these events were used as a user input into MODTRAN. The visibility values for these events were 9 km, 8 km, 7 km, 6 km, 4 km, 3 km, 1 km, 800 m, 600 m, 500 m, 100 m, 80 m, 50 m, and 10 m.
Figure 13 shows the atmospheric thermal radiation spectral distribution for some of the selected visibilities. For visibilities greater than 3 km, no considerable changes were found in the atmospheric thermal radiation spectral distribution for most wavelengths, and the spectral distribution resembles that of the clear sky as shown in Figure 1. For visibilities less than 1 km, the atmospheric thermal radiation spectral distribution has two features. First, the atmospheric emission outside the AW does not change greatly and remains fixed. In this case, the effect of the dust storm on the thermal spectral distribution is negligible. Second, the change of the spectral distribution inside the AW region is evident as we move toward lower visibilities (increasing the amount of aerosol). The increase in atmospheric thermal radiation continues up to visibility = 600 m. A dramatic increase in the atmospheric radiation in the AW region occurs for visibility of 10 m. For instance, as visibility reduced from 6 km to 80 m, the increase in the atmospheric thermal radiation in the AW region was approximately 210%. With 10 m visibility, the AW is totally closed, and the thermal emission resembles that of a BB. Therefore, the main effect of a dust storm on the atmospheric thermal radiation was to increase the thermal emission significantly inside the AW, causing the atmosphere to emit as a full BB.
Figure 14 shows the variations of the sky temperatures (integrated temperatures in the wavelength range of 5.5–50 μm) with the visibilities for 14 events. In this example, the sky temperature increases by 2°C as the visibility reduces from 9 to 3 km. At these visibilities, the effect of other meteorological parameters such as water vapor content, air temperature, and/or suspended aerosol particles are dominant. However, for visibilities below 1 km, the sky temperature increases dramatically with decreasing visibility. For example, at a visibility of 800 m, the sky temperature was −14°C, while at a visibility of 400 m, the sky temperature was −8°C, which is equivalent to an increase of approximately 6°C. For a visibility of 10 m, the integrated sky temperature was 4°C. This is approximately 37°C warmer than the clear atmosphere with a visibility of 9 km.
 Extensive measurements of the IR sky temperature were obtained using an IR detector operated at 5–50 μm for a period of more than 2 years in Riyadh, Saudi Arabia. Related ground level meteorological variables and radiosonde measurements were also recorded within the period. To study the effect of different atmospheric conditions on the sky temperatures, the data were divided into three different categories: clear, cloudy, and dusty sky conditions. The observation and research results are as follows.
 1. On clear-sky days, the sky temperature varied between −48.0°C and −2.1°C with a mean of about −28.2°C. We found that this wide range of temperatures is mainly due to the changes of atmospheric water content, screen level temperature and suspended aerosol particles in the atmosphere.
 2. On cloudy days, the sky temperature was between −37°C and 5°C. The mean sky temperatures were −10.2°C, −14°C and −17.5°C for total overcast, partly overcast, and scattered cloudy skies, respectively. The radiative properties of the sky in this case depend on a set of parameters related to the cloud characteristics (e.g., height, type and droplet size) and the intervening atmosphere between the ground and the cloud base. The influence of clouds on atmospheric thermal radiation decreases in importance with altitude because higher clouds are usually colder than lower clouds. The cloudy conditions warm up the atmosphere from 11°C to 18°C.
 3. For dusty sky conditions, investigations of the distribution of dusty events during the study period showed that the highest rate of dusty event occurred in March. The sky temperature in dusty conditions ranged between −20°C and 8.5°C. The mean sky temperatures for Blowing Dust, Dust Storm, and Severe dust Storm were −11.5°C, −6.2°C, and 2.9°C, respectively. The wide ranges of sky temperatures in these conditions is attributed to several factors, including the properties of the air masses that brought the dust and the dust particle characteristics such as size, shape, and chemical composition, which are initially determined by the sources from which the dust originated. The study showed that the dust or sand storm conditions increase the atmospheric temperatures by approximately 17–31°C, which is higher than that caused by cloudy conditions.
 4. Theoretical simulations using the MODTRAN software were conducted to investigate the atmospheric thermal radiation spectral distribution under different atmospheric conditions. The results showed that the major changes in the atmospheric spectra occurred mainly within the atmospheric window (8–14 μm). For instance, the atmospheric window is closed, and the atmospheric emission resembles that of the blackbody when the amount of water content reached its maximum. This phenomenon occurs in clear sky, in sky with low-level clouds and in strong or severe dust storm condition, when the visibility drops below 100 m. Moreover, little variations were observed outside the AW during such extreme conditions. The sky temperatures obtained from MODTRAN using the radiosonde data for Riyadh showed good agreements with the measured temperatures, which support the above conclusions that are derived from experimental data.
 The author would like to thank King Abdulaziz City for Science and Technology (KACST) for supporting this work. The Presidency of Meteorology and Environment (PME) is also acknowledged for providing the meteorological data. Special thanks go to Hamoud Al-Harbi, the director of the center, for his great support and encouragement. We also would like to thank Bader Al-harbi for providing us with dust storm photos. We also would like to thank the three reviewers for their valuable comments and suggestions.