2020 Comparison of Farmsense and sticky traps when monitoring with pheromone, PPO, and ovibait
~/ChuckBV/ms-now-pseudoaccounstic-sensors-2020/README.md
Overview
The objective of the current study is to examine correlation of trap activity as determined by FarmSense monitors when using pheromone, PPO, and pistachio ovibait as bait. The study used an RBCD arrangement, with sections of row-ends in a pistachio plot serving as replicate blocks.
The experiment is described in greater detail in "report-no-endnote_v02.pdf". A visual representation of the plot is found in "farmsense_experiment_2020_06_30b.jpg", and the Google Earth file "mwoolf-farmsense-expt-2020-07-01.kml" provides latitude/longitdue coordinates.
The Pheromone used has been Trece L2-l, and the PPO used has been the non-commercial pouches previously provided by a benefactor.
Replicate blocks were arranged along the east edge of the part of the Mike Woolf pistachio block stradling I-5 near Huron (this test was on the west of I-5) (see diagram in "./doc/"). The traps were placed in groups of groups of 6 serving as replicate blocks, with 7 such repliate blocks present. On various dates, traps were re-randomized.
Key dates
All wing traps and farmsense devices were in place by 9 July 2020. Traps were serviced weekly until 23 September 2020, at which time they were removed by owner request to avoid interference with harvest activities. Trapping resumed on 14 October 2020 for two more weekly trapping intervals, and traps were again removed on 4 November 2020.Monitoring intervals are summarized by script 1; see "./data/intervals.csv"
Data
./data-raw contains three csv data sets. "y20-farmsense-randomizations.csv" is a listing of trap positions within replicate blocks for replicates in time, which are distinguished by StartDate. "y20_farmsense_wingtrap_counts.csv" contains weekly count data for traditional wing traps with a glue liner. "captures_2020-07-01_2020-11-20_1605756009922.csv" has one record with time stamp and monitor identifier for each FarmSense detection event. These data were downloaded from the FarmSense server.
./data-intermediate contains five csv data sets. "intervals.csv" is a listing of start and end dates for weekly monitoring periods, obtained from a summary of "y20_farmsense_wingtrap_counts.csv". "data_skeleton.csv" was obtained by merging the randomizations and intervals data set, and serves as a scafold to merge wing trap and farmsense data into common data sets. "sticky_trap_counts_by_interval.csv" and "fsense_by_interval.csv" are the results of such merges. They have the same data struction and can be merged vertically {"rbind()"} into a common data set to facilitate comparisons. "fsense_all_obs.csv" replicate data but retains individual events to allow comparisons of the hour that the events were observed.
Script functions and output
script1a creates the data skeleton, while script1b and script1c merge on wing trap and FarmSense data, respectively. script1c outputs both the "fsense_by_interval.csv" and the "fsense_all_obs.csv" data sets.
script4 compares sticky trap counts over monitoring periods with sums of FarmSense sensor events per period at the level of individaul traps or sensors. sript4 output includes Figure 2 (a 3 x 1 grid of weekly mean and standard error of counts for traps or sensors baited with pheromone, ovibait, or PPO) and Figure 3 (scatter plots of Farm Sense events vs. trap counts for the same replicate and week). Statistics output byt script4 include a
- unequal variance t-test for ovibait
- Spearman correlation for pheromone
- Spearman correlatio for PPO
script5 produces vertical bar charts of counts by hour of the night in July, Aug, and Sep for ovibait, pheromone, and PPO