reads merged (raw+tx for v2+v3 stations) aws-l2 data
runs toL3 on that L2 file within which the GPS latitude, longitude and altitude are post-processed into smooth, gap-free series.
This is done by:
finding breaks in the available GPS coordinates which correspond to station maintenance. The smoothing and interpolation will be done separately for each interval between maintenance so that station relocation is not smoothed.
For each interval, we apply 1) a LOWESS (Locally Weighted Scatterplot Smoothing) to the each coordinate; 2) an internal linear interpolation between the available smoothed coordinates and 3) a linear extrapolation of the result to the next maintenance, to the latest timestamp or backward in time if necessary.
For stations where no GPS data is available, the static value contained as an attribute in the aws-l2 data file is given to each timestamp
The result are three new variables:
lat, lon, alt the smoothed and gapless time-dependent coordinates for each stations
And three new attributes:
lat_avg, lon_avg, alt_avg the mean value of the previous three variables that can be given as attribute in thre netcdf files and potentially in the AWS_metadata.csv file distributed with the dataset.
This PR builds on top of #262. It:
toL3on that L2 file within which the GPS latitude, longitude and altitude are post-processed into smooth, gap-free series.This is done by:
finding breaks in the available GPS coordinates which correspond to station maintenance. The smoothing and interpolation will be done separately for each interval between maintenance so that station relocation is not smoothed.
For each interval, we apply 1) a LOWESS (Locally Weighted Scatterplot Smoothing) to the each coordinate; 2) an internal linear interpolation between the available smoothed coordinates and 3) a linear extrapolation of the result to the next maintenance, to the latest timestamp or backward in time if necessary.
For stations where no GPS data is available, the static value contained as an attribute in the aws-l2 data file is given to each timestamp
The result are three new variables:
lat,lon,altthe smoothed and gapless time-dependent coordinates for each stations And three new attributes:lat_avg,lon_avg,alt_avgthe mean value of the previous three variables that can be given as attribute in thre netcdf files and potentially in theAWS_metadata.csvfile distributed with the dataset.