Snow Cover in Under Canopy and Open Locations at the HJ Andrews (OR, USA) and the Sagehen (CA, USA) Experimental Forests via Distributed Ground Temperature Sensors (Part I: July 2021 to July 2023)

These data are continued by the following data set:

Kostadinov, T. S. (2026). Snow Cover in Under Canopy and Open Locations at HJ Andrews (OR, USA) and Sagehen (CA, USA) Experimental Forests via Distributed Ground Temperature Sensors (Part II: July 2023 to July 2025) [Dataset]. Zenodo. https://doi.org/10.5281/zenodo.22101071

These data represent snow cover (expressed as binary snow presence or snow absence), determined using temperature sensors/loggers deployed on the ground in various under canopy and open (forest gap/meadow) locations at two coniferous montane forest sites in the Western USA: the HJ Andrews Experimental Forest (HJA) in Oregon, and the Sagehen Experimental Forest (Sagehen) in California. Measurements reported here were taken between July 2021 and July 2023.

At each sensor location, the following is provided: 

Snow classification for each date, including an assessment of classification uncertainty using a Monte Carlo simulation. A given day (in the PST time zone, i.e. UTC -8:00 hours) is classified as snow covered (a '1' in the data) if the absolute value of the mean temperature for the day is less than or equal to 1.0 deg. C and the standard deviation of the temperature is less than or equal to 0.353 deg. C,  otherwise it is classified as snow-free ('0' in the data). See the included code for details. For the criteria used, also see Kostadinov et al. (2019, doi: https://doi.org/10.1016/j.rse.2018.11.037) and Raleigh et al. (2013; doi: https://doi.org/10.1016/j.rse.2012.09.016).

Importantly, the canopy cover status of each sensor is provided as being either under canopy or in the open. This was determined manually/visually in the field at deployment by ascertaining canopy cover or its absence vertically directly above each sensor. 

Raw temperature files are also included, as output by the Onset(R) HOBOware(R) Pro software. Date/time has been re-formatted in these raw temperature files, and the column headings might be redacted, but otherwise the data is as output by the sensor and HOBOware(R) Pro. 

Calibration in an ice slush was conducted for a subset of the sensors prior to deployment, and these calibration data are included as well (Important Note: the calibration values are not applied to the data set).

The data are given in CSV format. Figures of the temperature, snow classification and calibration data are provided for convenience as well. 

Data analysis and processing, snow classification, as well as figure production was accomplished in MATLAB(R), and the relevant scientific code is included here. Both raw and snow classification files given here were processed and output by MATLAB(R). 

Geographic location for each sensor was determined in the field with a Juniper Systems(TM) Geode (TM) GNS2 receiver, using the ESRI(R) ArcGIS(R) Collector app to collect the locations and average 40 points at each sensor location for improved accuracy, typically sub-meter, enabling high-resolution applications such as tree-scale and lidar studies. For both HJA and Sagehen, a separate CSV file summarizing location, location uncertainty and canopy cover metadata for each sensor is included. Some key parts of these data are also included in each sensor's file headers and in the figures.