How Solar Radiation Was Estimated
Routine monitoring of solar radiation is conducted at relatively few locations in Hawai‘i. While measurements are very valuable, the number and distribution of stations is not sufficient to enable us to map solar radiation patterns based only on station data. Therefore, we have to make estimates that combine modeling and observations from ground stations and satellites. We mapped solar radiation by first estimating clear-sky solar radiation, the amount of sunlight received with no clouds at a given location, time of year, and time of day. Then the effects of clouds and shading by surrounding terrain were incorporated to produce maps of solar radiation. We analyzed patterns of cloud frequency based on imagery from the MODIS and GOES satellite platforms.
Clear-Sky Solar Radiation
We used the REST2 clear-sky radiation model (Gueymard, 2008) to estimate the pattern of solar radiation without clouds. REST2 accounts for variations in sun angle and allows for adjustment of atmospheric optical properties to account for absorption and reflection of radiation as it passes through a cloud-free atmosphere.
Clouds are a major factor in determining patterns of solar radiation in Hawai‘i. We used MODIS and GOES satellite imagery to estimate the patterns of cloud frequency. MODIS has a spatial resolution of 1 km, but takes images over Hawai‘i only four times per day (only twice during daylight hours). GOES, on the other hand, has a coarser spatial resolution (4 km), but takes images every 15 minutes. In this project, we devised a way to combine cloud data from these two sources and take advantage of the high spatial resolution of the MODIS data and the high temporal resolution of the GOES data. Fusing them together, we produced maps of cloud frequency at 1 km resolution for each hour of the average diurnal cycle of each month, i.e., 24 x 12 maps were produced. These maps were then resampled to a higher resolution (250 m).
Linking Cloud Frequency to Solar Radiation
To incorporate the effects of clouds in the estimates of solar radiation, we used observations of solar radiation at ground stations to derive a statistical function relating cloud frequency with the reduction of solar radiation by clouds. With this cloud function, the cloud frequency maps could be applied to the clear-sky solar radiation maps to derive maps of solar radiation.
In mountainous areas, solar radiation can be affected when the surrounding terrain blocks the sun. To include this effect we determined the shaded map pixels for each hour and month based on the sun angle in relation to relative elevation of terrain surrounding each location. A pixel was designated as shaded whenever the sun’s elevation angle was less than the local horizon elevation angle. For shaded pixels, we assumed that the direct radiation was zero and the diffuse radiation was unaffected (total or “global” solar radiation is the sum of direct beam radiation and diffuse radiation; diffuse radiation is the portion of solar radiation that reaches the surface after being reflected by atmospheric gases, aerosols, clouds, or surrounding terrain). We estimate diffuse radiation as a function of cloud frequency, and set solar radiation equal to the diffuse portion for shaded pixels.