solar_irradiation_tests
I downloaded monthly solar irradiation values from the RAWS USA Climate Archive for three weather stations (# months, start-end dates) nearest the Kaibab and Raplee field sites:
Paria Point (265, 9/1994-8/2016), House Rock (179, 3/1999 - 11/2013), and Kane Gulch (222, 3/1998-8/2016).
Next, I selected a (single pixel) region from the Raplee Ridge surface which experienced full daily sunlight without the influence of local topography (i.e. the top of the ridge).
I used the point sampling tool in QGIS to sample all twelve months of the Raplee Ridge data.
In R I imported the four CSVs and plotted the three weather stations as points, and the Raplee Ridge as lines using ggplot2.
I compare the impact of three cloud scenarios for the Raplee Ridge:
(1) The simple clear sky r.sun models from Comet using default settings: (global_radiation)
(2) The clear sky models with cloudiness factored from the MODIS derived monthly cloud layer (Wilson and Jetz 2016): (global_radiation * (1-cloudiness))
and
(3) The clear sky models with 1/2 the cloudiness from Wilson and Jetz (2016): (global_radiation * (1-cloudiness/2)).
The Wilson and Jetz (2016) cloud layer units are scaled between 0.01 and 0.99 for 1% and 99% cloudiness, respectively.
raplee <- read.csv('f:/raplee_kaibab/raplee_high_point_solar.csv')
paria <- read.csv('f:/raplee_kaibab/paria_point_solar.csv')
kane_gulch <- read.csv('f:/raplee_kaibab/kane_gulch_solar.csv')
houserock <- read.csv('f:/raplee_kaibab/houserock_solar.csv')
library(ggplot2)
p <- ggplot(data=houserock,aes(x=month,y=w_m2))+geom_line(data=raplee, aes(x=month,y=w_m2),color = "black")+geom_line(data=raplee, aes(x=month, y=w_m2*(1-cloud)))+geom_line(data=raplee, aes(x=month, y=w_m2*(1-cloud/2)))+geom_point(data=kane_gulch,aes(x=month,y=w_m2))+geom_point(data=paria,aes(x=month,y=w_m2))
p
Figure: The range of monthly solar irradiation for the three sites are combined as box plots (red line is the median, blue box is 25th to 75th percentile, whiskers are +/–2.7σ and 99.3%, and outliers are crosses). The three black lines plotted through the graph represent in descending height (a) clear sky, (b) 1/2 cloud, and (c) full cloudiness. There are a total of 666 months in the dataset with at least 50 months for each month from which the box plot distribution is generated.
In general the r.sun Raplee Ridge global_radiation model with no cloudiness is consistent with the brightest months recorded at the three stations; while the full cloudiness is consistent with the least bright months recorded at the three stations.
Note: All three stations are at different elevations and aligned with different landform positions which are unmeasured. These may influence shading as well as daily local orographic formation of clouds, as reported by Wilson and Jetz (2016).