kayad.alfalfa | R Documentation |
Yield monitor data for 4 cuttings of alfalfa in Saudi Arabia.
data("kayad.alfalfa")
A data frame with 8628 observations on the following 4 variables.
harvest
harvest number
lat
latitude
long
longitude
yield
yield, tons/ha
Data was collected from a 23.5 ha field of alfalfa in Saudia Arabia. The field was harvested four consecutive times (H8 = 5 Dec 2013, H9 = 16 Feb 2014, H10 = 2 Apr 2014, H11 = 6 May 2014). Data were collected using a geo-referenced yield monitor. Supporting information contains yield monitor data for 4 hay harvests on a center-pivot field.
# TODO: Normalize the yields for each harvest, then average together # to create a productivity map. Two ways to normalize: # Normalize to 0-100: ((mapValue - min) * 100) / (max - min) # Standardize: ((mapValue - mean) / stdev) * 100
Ahmed G. Kayad, et al. (2016). Assessing the Spatial Variability of Alfalfa Yield Using Satellite Imagery and Ground-Based Data. PLOS One, 11(6). https://doi.org/10.1371/journal.pone.0157166
None
library(agridat)
data(kayad.alfalfa)
dat <- kayad.alfalfa
# match Kayad table 1 stats
libs(dplyr)
dat <- group_by(dat, harvest)
summarize(dat, min=min(yield), max=max(yield),
mean=mean(yield), stdev=sd(yield), var=var(yield))
# Figure 4 of Kayad
libs(latticeExtra)
catcols <- c("#cccccc","#ff0000","#ffff00","#55ff00","#0070ff","#c500ff","#73004c")
levelplot(yield ~ long*lat |harvest, dat,
aspect=1, at = c(0,2,3,4,5,6,7,10), col.regions=catcols,
main="kayad.alfalfa",
prepanel=prepanel.default.xyplot,
panel=panel.levelplot.points)
# Similar to Kayad fig 5.
## levelplot(yield ~ long*lat |harvest, dat,
## prepanel=prepanel.default.xyplot,
## panel=panel.levelplot.points,
## col.regions=pals::brewer.reds)
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.