search

vscholl / neonVegWrangleR

Wrangling NEON vegetation structure (vst) for integration with Airborne Observation Platform (AOP) remote sensing data
Other
1 stars 5 forks source link

Notes on retrieve_VST_data #5

Open bw4sz opened 4 years ago

bw4sz commented 4 years ago

Hey @vscholl, i'm the postdoc in @MarconiS lab. I'm going to debug some errors i'm seeing in the field data retrieve function. Its not clear to me if its an error of code here or with NEON. I'll be added a few issues to keep track of my thoughts. I'm working from Sergio's dev branch, I forked from there.

  1. Describe param args. I assume this is in error. https://github.com/vscholl/neonVegWrangleR/blob/71c7469c9d85a3f9b57e58cfeee5ec98bb902a5f/R/retrieve_VST_data.R#L4

I took a shot at it, but i'll need feedback. 

#' @param site (character) "all" for all sites (see neonUtilities::loadByProduct), else a NEON site ID
#' @param start (character) "Either NA, meaning all available dates, or a character vector in the form YYYY-MM, e.g. 2017-01. Defaults to NA"
#' @param enddate Either NA, meaning all available dates, or a character vector in the form YYYY-MM, e.g. 2017-01. Defaults to NA.
MarconiS commented 4 years ago

basically that is the option in which we use only geoNEON approach. That is fine, but is quite slow and for a few points is very imprecise (more than 400m off). The parameter are fine, at least they are consistent to neonUtilities package descriptions. Is the error you referring to in the eventIDs?

bw4sz commented 4 years ago

I will open up multiple issues with tests that fail. Here is one upstream of this (i think).

https://github.com/vscholl/neonVegWrangleR/issues/6

I haven't made it to the original error yet.

MarconiS commented 4 years ago

I think I got where the problem is embedded. There are two different EventID, one from "mappingAndTagging", where we calculate the coordinates for each individual tree crown, one for "vst_apparentindividual" (where biometric measures are taken). These two don't necessarily match. We can either rename mappingEventID, but that would likely increase the number of repeated biometric measures ((which we don't want), or just keep the EventID from apparent_individuals (but we lose the information of when coords measures have been collected). either way works fine to me