Comparing met data from various sources

Sources:

ebifarm local met station Data
- point location (40.08\( ^o \) N, -88.2\( ^o \) W), energy farm, Urbana, IL
narr [NARR daily](http://www.esrl.noaa.gov/psd/data/gridded/data.narr.html]
- 0.3\( ^o \) grid daily 1979-2013
narr3h NARR from MsTMIP
- 0.25\( ^o \) grid 3-hourly 1979-2010
cruncep CRU-NCEP from MsTMIP
- 0.5\( ^o \) grid 6 hourly 1901-2010

Variables

RH / Relative Humidity (%)
PAR / Solar Radiation (umol/h/m2)
Temperature ©
Wind Speed (m/s)
Precipitation (mm/h)

Note there is obviously an error in the downscaling method used to generate hourly NARR from daily. This is an obsolete method (since we now have sub-daily, and if we are going to correct this, it should be in the version of weachDT currently in PEcAn.data.atmosphere)

Comparing 'data' (ebifarm) with gridded products

TODO: clean up figure titles, labels, write explanations

library(PEcAn.data.atmosphere)

## Loading required package: ncdf4
## Loading required package: udunits2
## Loading required package: ggplot2

theme_set(theme_bw())
data(narr_cruncep_ebifarm)

knitr::opts_chunk$set(echo = FALSE, cache = TRUE, comment = NA, tidy = TRUE, 
    warning = FALSE, results = "hide", fig.width = 10, fig.height = 4)

Extracting data

These data are on biocluster.igb.illinois.edu, most 10-100s GB. Scripts used to download and convert these data to PEcAn CF format, optimized for time series extraction, are on GitHub ebimodeling/model-drivers.

Solar Radiation (PAR) vs Temp

plot of chunk solar-v-temp

RH vs Temp

plot of chunk RH-v-Temp

Solar Radiation and Precipitation: NARR daily vs 3 hourly

plot of chunk par-v-precip

Precipitation v Temperature

plot of chunk precip-v-temp

Compare Solar Radiation

plot of chunk solar

Max Solar Radiation for June 1-Aug31 2010

plot of chunk max-solar-plot

Max Solar Radiation (PAR) Model v OBS

plot of chunk create-plots

PAR residuals (model - obs)

geom_smooth: method="auto" and size of largest group is >=1000, so using gam with formula: y ~ s(x, bs = "cs"). Use 'method = x' to change the smoothing method.
geom_smooth: method="auto" and size of largest group is >=1000, so using gam with formula: y ~ s(x, bs = "cs"). Use 'method = x' to change the smoothing method.
geom_smooth: method="auto" and size of largest group is >=1000, so using gam with formula: y ~ s(x, bs = "cs"). Use 'method = x' to change the smoothing method.

plot of chunk solarresid-plot

Correlations of daily max solar radiation

plot of chunk maxsolar-plot

Compare daily and 3hourly downscaled NARR

plot of chunk unnamed-chunk-1

Multiple variables

plot of chunk all-vars-plots

Some sanity checks on variables

Temperature:

|source   |      min|   mean|    max|
|:--------|--------:|------:|------:|
|cruncep  |  -19.582|  10.77|  30.54|
|narr     |   -5.098|  12.60|  24.19|
|narr3h   |  -17.047|  12.46|  36.24|
|ebifarm  |  -23.770|  10.91|  34.22|

|source   |     min|   mean|     max|
|:--------|-------:|------:|-------:|
|cruncep  |  15.959|  71.78|  100.00|
|narr     |  47.650|  84.35|   99.21|
|narr3h   |   5.902|  68.84|  100.00|
|ebifarm  |  13.590|  70.64|   95.70|

Total Precip

|source   |   total|
|:--------|-------:|
|cruncep  |   947.4|
|narr     |  1108.5|
|narr3h   |   772.0|
|ebifarm  |   385.1|

More diagnostic plots

need to print each one …

geom_smooth: method="auto" and size of largest group is >=1000, so using gam with formula: y ~ s(x, bs = "cs"). Use 'method = x' to change the smoothing method.
geom_smooth: method="auto" and size of largest group is >=1000, so using gam with formula: y ~ s(x, bs = "cs"). Use 'method = x' to change the smoothing method.
geom_smooth: method="auto" and size of largest group is >=1000, so using gam with formula: y ~ s(x, bs = "cs"). Use 'method = x' to change the smoothing method.
geom_smooth: method="auto" and size of largest group is >=1000, so using gam with formula: y ~ s(x, bs = "cs"). Use 'method = x' to change the smoothing method.
geom_smooth: method="auto" and size of largest group is <1000, so using loess. Use 'method = x' to change the smoothing method.

plot of chunk more-plots

geom_smooth: method="auto" and size of largest group is >=1000, so using gam with formula: y ~ s(x, bs = "cs"). Use 'method = x' to change the smoothing method.
geom_smooth: method="auto" and size of largest group is >=1000, so using gam with formula: y ~ s(x, bs = "cs"). Use 'method = x' to change the smoothing method.
geom_smooth: method="auto" and size of largest group is >=1000, so using gam with formula: y ~ s(x, bs = "cs"). Use 'method = x' to change the smoothing method.
geom_smooth: method="auto" and size of largest group is >=1000, so using gam with formula: y ~ s(x, bs = "cs"). Use 'method = x' to change the smoothing method.
geom_smooth: method="auto" and size of largest group is <1000, so using loess. Use 'method = x' to change the smoothing method.

plot of chunk more-plots