4.1. Temporal Variability of Cloudiness and Solar Energy Attenuation in San Antonio, Texas
shows the frequency of cloud types and the corresponding reduced solar energy (both in percentage) calculated for each year in 2009–2016 using GSIP-v2 and GSIP-v3 data for the UTSA and ASF locations. The total number of observations for each year (see upper right corner of each panel) is ~7000 except for 2009 and 2016, which are smaller (~6000) as expected, since both are incomplete years missing the first 3, and last ~2 months, respectively.
is similar to Figure 2
but combines all available data for GSIP-v2 and GSIP-v3. When comparing the results for GSIP-v2 and GSIP-v3 in Figure 2
and Figure 3
, especially when comparing the overlapping year 2014, it is apparent that GSIP-v2 gives smaller frequency occurrences of clear-sky conditions and greater frequency occurrences of cloudy-sky conditions. It is not clear why this is so, and to the best of our knowledge, there was no previous report that documented these discrepancies. The difference in spatial resolution (14 km for GSIP-v2 versus 4 km for GSIP-v3) could have potentially contributed to such discrepancy. GSIP-v3 has higher spatial resolution and is thus chosen for further analysis of cloud and solar energy potential in this study.
By looking at the 2014, 2015, 2016 plots from GSIP-v3 in Figure 2
, it is clear that the frequency of occurrence of cloud types and their attenuation effects are very similar from year to year, i.e., clear-sky conditions are the most common and they cause the least reduction of solar irradiance, and the glaciated clouds and multilayered clouds are always the least common but cause the greater reduction of solar irradiance. The influence of the different cloud types on solar irradiance is found to be the same as expected from [20
]. As seen in the GSIP-v3 plot in Figure 3
(right panel), clear skies in both San Antonio locations occur ~57% of the time and reduce solar irradiance by ~33%; cirrus clouds occur ~17% of the time and reduce solar irradiance by ~56%; water clouds occur ~15% of the time and reduce solar irradiance by ~50%; and the other cloud types (mixed, partly, multilayered and glaciated) account for less than 5% of sky conditions and could reduce solar irradiance by 57–75%.
In Figure 4
, the diurnal (left panel) and seasonal (right panel) variabilities of cloud properties and their effect on attenuating irradiance are presented. To investigate the daily variations (color bars in Figure 4
, left), the frequency of occurrence of cloud types and their respective irradiance attenuation are stratified into four separate even 2.25-h periods: 07:45–10:00 (early morning), 10:45–13:00 (late morning), 13:45–16:00 (early afternoon), and 16:45–19:00 (late afternoon). Overall, the frequency of occurrence of each cloud type does not change much during the day, with the frequency of the clear-sky conditions showing a slight increase from early morning to late afternoon, slightly more noticeable at the UTSA location. The late afternoon (16:45–19:00) has 4–7% more clear-sky conditions than in the early morning (07:45–10:00). Overall, partly and mixed clouds occur more frequently in the early morning than in any other time period. Water, glaciated, and cirrus clouds occur more frequently in the late morning and early afternoon, while the mixed and multi-layered clouds occur more in the early morning and late afternoon.
In terms of the daily variations in solar irradiance attenuation for each cloud type (color error bars in left panel of Figure 4
), overall, the attenuation of solar irradiance in percentage is higher in early morning and late afternoon (when the solar zenith angle is larger), with some exceptions. The reduced solar irradiance in percentage under clear-sky conditions is higher with 39–44% in the early morning and late afternoon and lower with 26–28% in the middle of daytime, most likely due to the longer path that solar radiation goes through the atmosphere in the early morning and late afternoon. Under partly, water, mixed, cirrus, and multilayered clouds, the reduced solar irradiance in percentage is the highest in the early morning and the lowest in the early afternoon. The reduced solar irradiance under glaciated clouds is the lowest with ~67% in the late afternoon than any other time periods.
To investigate the seasonal variations (Figure 4
, right), the frequency of occurrence of cloud types and their respective irradiance attenuation are stratified into four separate three-month periods: Winter (December-February), spring (March–May), summer (June-August), and fall (September-November). The frequencies of clear-sky occurrences are the highest in summer (69%) and the lowest in spring (43%). Water and cirrus clouds are the most common in spring (19–22%) and the least common in summer (11–12%). Partly and mixed clouds are more common in spring and winter but less common in summer and fall. Glaciated and multi-layered clouds are more common in spring than in any other seasons. Overall, the attenuation of solar irradiance for each cloud type varies only slightly from season to season.
To investigate the daily variations in each season, Figure 5
shows the frequency of occurrence and irradiance attenuation for each cloud type stratified as in Figure 4
(left) but for each season defined as in Figure 4
(right). The frequencies of occurrence of cloud types are different for each season. The frequency of occurrence of clear-sky conditions in spring is similar to the all-season average shown in Figure 4
(left), although the frequency of occurrence in spring is much smaller than the all-season averaged and also smaller compared with to the other three seasons. Summer has the highest clear-sky percentage of occurrence with even higher values in the late morning and early afternoon than in the two other time periods. Fall has the second highest clear-sky percentage, with the highest frequency in the early morning period. Winter ranks the fourth in terms of clear-sky frequency of occurrence, with the highest frequency in the late afternoon period. Note that the patterns of reduced solar irradiance in percentage by cloud types during each time period, among four seasons, are similar, but with slightly different magnitudes.
4.2. Spatial Distribution of Clouds and Solar Energy Potential in Texas and Nearby Regions
and Figure 7
show the seasonal distribution of cloud-type and cloud-layer frequencies for the four seasons over the study area. In Figure 6
, clear-sky conditions overall are more common than cloudy-sky conditions. Cirrus is the most common cloud type, followed by water clouds. The other four types (partly, mixed, glaciated, and multilayered) are less common (mostly less than 15%). All of these results are consistent with the findings for the two San Antonio locations presented in Section 4.1
. The frequency of clear sky conditions in winter is higher in the center and western parts of the study area, especially in the northern Mexico region; in spring, it is higher mostly in the western-most (middle latitudes) region of the study area and in the southeastern corner over the Gulf of Mexico; in summer, it is higher mainly in Texas and northeast region of Mexico, with higher frequencies also over the Gulf of Mexico; in fall, the frequency of clear sky is more homogeneous in space, except for some lower frequency spots in the Gulf of Mexico and near the Sierra Madre Occidental in Mexico. The Sierra Madre Occidental also has low frequency of clear sky in summer, perhaps due to topographic blocking of clouds and/or clouds formed over the range with warm moist air rising up.
The frequency of cirrus clouds is relatively higher in Gulf of Mexico and lower over land areas, except along the Sierra Madre Occidental where the cirrus clouds seem occur frequently in all seasons which is consistent with the lower frequency of clear-sky conditions in these region noted earlier for the summer but that it is apparent also in the other seasons in Figure 6
. Water clouds have higher frequencies in the Gulf of Mexico in winter, while this seems to occur more frequently over land than over the ocean in spring and summer, with a more homogeneous spatial distribution in the fall.
In terms of the overall spatial distribution of cloud-layer frequencies (Figure 7
), low clouds occur more frequently. Around Texas, the frequencies of low clouds are higher in winter and lower in spring. In the Gulf of Mexico, low clouds are more frequent in summer and less frequent in winter. Mid altitude clouds occur more frequently in winter/fall than in the other seasons over the Gulf of Mexico. Over land, high clouds occur more frequently in spring and over the ocean in spring/fall. The Sierra Madre Occidental shows higher frequency of mid- and high-clouds, but less frequency of low clouds than other areas.
The monthly spatial distribution of solar energy potential is presented in Figure 8
. Depending on the calendar month and location, the monthly solar energy potential ranges from 43–254 kWhm−2
. from November to January, the solar energy in the north to northeast part of the study area is quite low (<100 kWhm−2
). It is higher in parts of Mexico and the Gulf of Mexico, but less than 160 kWhm−2
. In February, the solar energy potential increases, especially over Mexico and the Gulf of Mexico, with the increases of 10-30 kWhm−2
in the central region of the study area. In March, the solar energy in the southwestern and southeastern corners of the study area reach values of 190–210 kWhm−2
. In this month, the northwestern Texas region has a monthly solar energy potential of 150–180 kWhm−2
and the northeastern Texas region still has low values. From April to July, the available solar energy continues to increase to reach maximum values over most of the study area. An exception is the Sierra Madre Occidental, where the energy potential reaches maximum values in May. In August, the solar energy potential over most of the study area starts to drop to 180–210 kWhm−2
over Texas. The available solar energy over Gulf of Mexico drops by 10–20%, but is still higher than that any region over land. In September, the distribution of solar energy seems relatively homogeneous, with overall values of 140–200 kWhm-2
. The Gulf of Mexico receives significantly less solar energy in September than in any month in the period March-August. In October, the solar energy potential in the upper part of the study area generally drops to 100–150 kWhm−2
, while it ranges from 150–180 kWhm−2
over Gulf of Mexico.
The available seasonal solar energy potential ranges from 165-710 kWhm−2
) with maximum vales in summer and minimum in winter. In winter, the seasonal solar energy potential does not exceed 350 kWhm−2
over the most of Texas and 420 kWhm−2
over Gulf of Mexico. In spring, the seasonal energy potential increases to more than 500 kWhm−2
over Texas with parts of Mexico and the Gulf of Mexico receiving more than 600 kWhm−2
. The highest amount of solar energy is received in summer over the entire study area, particularly in the Gulf of Mexico. In fall, the seasonal energy potential drops by 300–450 kWhm−2
over the entire study area.
The annual solar energy, shown in the bottom right panel of Figure 9
, is derived only for the year 2015, because it is the only complete year in the GSIP-v3 dataset. The annual energy potential ranges from 1300–2300 kWhm−2
. The northeastern region of the study area receives lower values of solar energy (less than 1800 kWhm−2
). The north-western part of Texas receives 50–200 kWhm−2
more energy than the rest of Texas. The Gulf of Mexico and Mexico receive more solar energy than Texas and the north-eastern region of the study area.
4.3. Time Series of Solar Energy Potential in San Antonio, Texas
As an example of the daily, monthly, and seasonal solar energy potential time series derived from the satellite data, the time series at UTSA and ASF are presented in Figure 10
. For both UTSA and ASF, it is found that the highest daily solar energy potential is ~8.58 kWhm−2
on June 10, 2014 but the lowest is 1.16 kWhm−2
on December 17, 2014. The min, mean, and max values of daily solar energy in each month are also presented. The monthly solar energy varies from 79 to 217 kWhm−2
. from October to February, the monthly solar energy does not exceed 120 kWhm−2
, while it is consistently above 145 kWh m−2
from March to September. The range of monthly solar energy potential reaches 155–200 kWh m−2
in April, May and June. In July, the range of solar energy potential is close to 210–225 kWh m−2
. The solar energy potential drops to 190–200 kWhm−2
in September. The seasonal solar energy potential reaches the highest values 595–615 kWhm−2
in summer and the lowest values 335–340 kWhm−2
in winter. In fall, it reaches values of 395–410 kWhm−2
. The received solar energy in spring is around 485–510 kWhm−2
. The annual solar energy potential is 1750 kWhm−2
. A general good agreement is found between satellite-derived solar energy potential and the estimated energy potential from ground observations at the UTSA and ASF stations, as expected from Xia et al. [14
] who found that the global horizontal irradiances from satellite and ground at the two stations also agreed well (correlations 0.80–0.87 on the hourly timescale and 0.94–0.91 on the daily timescale). In most cases, the satellite-derived monthly solar energy potential is higher than that derived from the ground observations, except in February and July. The highest difference in monthly solar energy potential is observed in May, followed by September. For the seasonal energy potential the larger difference between satellite and ground estimates is found in spring.