PointSnobal

Python wrapped implementation of the Snobal model applied at a point.

The code in pointsnobal/c_snobal is the same underlying algorithms described in A spatially distributed energy balance snowmelt model for application in mountain basins (Marks 1999), which details Snobal. This code was originally available in IPW.

This software takes in a csv of HOURLY input data and writes a csv of daily snowpack data.

Versions

PointSnobal ships in versioned lines:

Research API

🚀 Calling Snow Researchers & Students! 🚀

At M3 Works, we’re passionate about advancing scientific research and education! 🌍📊

That’s why we offer free access to our PointSnobal API for qualifying research and educational projects. Whether you’re modeling snow processes or exploring hydrology, we’re here to support your work!

🔑 Request an API key on our website -> https://m3works.io/contact

Let’s collaborate and push the boundaries of environmental modeling together! ❄️📡

Credit

If you use the pointsnobal API we ask that you credit M3Works. To cite, please use the following DOI:

https://zenodo.org/records/14814813

Disclaimer

This API is provided under the same license listed in this directory. The API and code are provided “as is”, and M3 Works LLC makes no guarantees of functionality, performance, or fitness for any particular purpose.

M3 Works LLC shall not be held liable for any direct, indirect, incidental, or consequential damages arising from the use or inability to use this API or its associated code.

Use of this API constitutes acceptance of the terms and conditions outlined in the accompanying license.

API usage

Example of api usage in python

from pathlib import Path
import requests
import pandas as pd

api_key = "<your key>"
api_id = "bktiz24e19"
file_path = Path("<path to your input csv file>")
elevation = 1000  # your point elevation in meters (REQUIRED)
url = f"https://{api_id}.execute-api.us-west-2.amazonaws.com/m3works/snobal"
# Query parameters. Only `elevation` is required; the others are optional and
# shown here at their defaults. z_u / z_t / z_g are measurement heights in
# meters, relative to the snow surface.
params = {
    "elevation": elevation,       # REQUIRED - point elevation (m)
    "z_u": 5.0,                   # wind speed measurement height (m)
    "z_t": 2.0,                   # air temperature measurement height (m)
    "z_g": 0.3,                   # soil temperature depth (m)
    "output_frequency": "daily",  # "daily" (one row per day) or "hourly"
}

output_file_name = file_path.name.replace('inputs', 'snobal')
output_file = file_path.parent.joinpath(output_file_name)

# Headers
headers = {
    "x-api-key": api_key,
    "Content-Type": "text/csv"
}

print("Reading file and calling API")
# Load the CSV file as binary data
with open(str(file_path), "rb") as file:
    response = requests.post(
        url, headers=headers, params=params, data=file)

print("API request finished")
# error if we failed
response.raise_for_status()

result = response.json()

print("Parsing results")
# Get result into pandas
df = pd.DataFrame.from_dict(result['results']["data"])

Running locally

Script usage

Use scripts/use_api.py to call the api from the command line

python3 scripts/use_api.py <path to your file>  \
<your point elevation> --api_key <your api key>

This will output a csv file of the results. Run python3 scripts/use_api.py --help for a full list of options.

Input files

Variables that inform Snobal

These variables are directly used within Snobal

• air_temp - modeled air temp at 2m above ground
• percent_snow - % mix of snow vs rain (1 == all snow) [decimal percent]
• precip - precipitation mass [mm]
• precip_temp - wet bulb temperature
• snow_density - density of the NEW snow that hour [kg/m^3]
• vapor_pressure - modeled vapor pressure
• wind_speed - Wind speed at 5m above ground [m/s]
• soil_temp - Average temperature of the soil column down to 30cm
• net_solar - Net solar into the snowpack [w/m^2]
• thermal - Incoming longwave radiation into the snowpack [w/m^2]

See ./tests/data/inputs_csl_2023.csv for an example of data format

Measurement heights

Heights are relative to the snow surface and default to the values below. Override them per run with the z_u / z_t / z_g parameters — as API query parameters (above), or on the CLI with make_snow --z_u/--z_t/--z_g.

• z_u — wind speed measurement height (default 5 m)
• z_t — air temperature measurement height (default 2 m)
• z_g — soil temperature depth (default 0.3 m)

[!IMPORTANT] Watch out for...

• Snobal expects temperatures to be in Kelvin. This code expects Celsius. We do the conversion in get_timestep_force
• Precip mass (precip) is a big driver here. Without accurate conditions, model results will be poor

Local Install

[!TIP] Creating a local virtual environment with your tool of choice is recommended to isolate your code prior to installation.

Download Code

Navigate to a directory where you would like to download the repo, for example a projects directory in your home, and clone the repository.

cd ~/projects
git clone git@github.com:M3Works/pointsnobal.git
cd pointsnobal

Requirements

Requirements can be found in requirements.txt.

[!NOTE] A C-compiler with OpenMP support is required, on linux this is generally available. On macOS using Homebrew is a simple option.

brew install gcc libomp

For local build:

pip install -r requirements.txt
python3 setup.py build_ext --inplace
python3 install .

PointSnobal script

The entrypoint is make_snow once installed:

make_snow <path to input file> <elevation in meters>

For example, the installation can be quickly verified by running the test problem.

make_snow ./tests/data/inputs_csl_2023.csv 2101 --output_file test.csv

Validation data

Using metloom for station data that can be used for validation. get_daily_data returns a GeoPandas DataFrame of the variables and units on a daily timestep. Validation is crucial in snowpack modeling!

# Imports
import pandas as pd
from metloom.pointdata import CDECPointData
from metloom.variables import CdecStationVariables

# Specify start and end date
start_date = pd.to_datetime("2019-10-01")
end_date = pd.to_datetime("2020-06-01")

# List of variables to request
desired_variables = [
    CdecStationVariables.SWE, CdecStationVariables.SNOWDEPTH
]

# Define the point
point = CDECPointData("GRV", "Graveyard Meadow")

# Request the data
df = point.get_daily_data(
    start_date, end_date, desired_variables
)
# Data comes back indexed on `datetime` and `site`, reset to just datetime
df = df.reset_index().set_index("datetime")
# Show the results
print(df)
# store in csv if you want
df.to_csv(f"{point.id}_station_data.csv", index_label="datetime")

Troubleshoting

Install issues on macOS

If you are getting 'omp.h' file not found or ld: library not found for -lomp during the setup.py build_ext step you need to be sure that the correct compiler is being utilized and the OpenMP libraries are available.

First, set the CC environment variable to your C-compiler. For example, the following is are the paths used on macOS when using homebrew.

export CC=/usr/local/bin/gcc-14 # Intel
export CC=/opt/homebrew/bin/gcc-14 # Apple silicon

Second, be sure the OpenMP libraries are available to the compiler. This can be accomplished by setting the LDFLAGS envornment variable. For example, for libomp install with homebrew.

export LDFLAGS="-L/opt/homebrew/opt/libomp/lib" # Intel
export LDFLAGS="-L/opt/hombrew/Cellar/libomp/lib" # Apple silicon

For additional help on these path, when using homebrew, utilize the brew info gcc and brew info libompcommands.