LandBOSSE data format mismatch

[max-monere]

Describe the bug

When using a WIND Toolkit-formatted .csv weather file with the LandBOSSE BOS cost module, SAM returns the error: Run LandBOSSE error: Error in Weather_Data: Length of values does not match length of index. This occurs because the read_weather_data() function in landbosse_api/run.py hardcodes skiprows=5, assuming the input weather file is in .srw format with exactly 5 header rows. WIND Toolkit .csv files have only 2 header rows, so skiprows=5 consumes 3 rows of actual data, leaving fewer than 8,760 data rows. When LandBOSSE then attempts to assign a pre-built 8,760-entry datetime index to this shorter dataframe, pandas raises the length mismatch error.

To Reproduce

1. Download a hourly wind resource file from the NREL WIND Toolkit in .csv format
2. Open SAM and create a Wind Power project
3. Assign the WIND Toolkit .csv as the weather resource file
4. Enable the LandBOSSE BOS cost model and fill in required inputs
5. Run the simulation — the BOS calculation returns the error described above

Expected behavior

LandBOSSE should accept WIND Toolkit .csv formatted weather files without error, consistent with SAM's own documentation noting a transition toward WIND Toolkit .csv format compatibility. The read_weather_data() function should either detect the number of header rows dynamically, or SAM should preprocess the weather file into .srw format before passing it to LandBOSSE.

Screenshots

Operating System and Version

• OS: Windows
• Version: SAM 2025.4.16

Additional context

The relevant code is in landbosse_api/run.py (installed with LandBOSSE 2.3.0.3 in SAM's bundled Python environment at runtime/python/python-3.7.6/Lib/site-packages/landbosse/). The specific function is:

def read_weather_data(file_path):
    weather_data = pd.read_csv(file_path,
                               sep=",",
                               header=None,
                               skiprows=5,
                               usecols=[0, 1, 2, 3, 4]
                               )
    return weather_data

A workaround is to manually add 3 dummy header rows to the WIND Toolkit .csv file (bringing the total to 5 header rows) before using it with LandBOSSE. This modified file will break SAM's main production simulation, so two separate copies of the weather file are needed — one unmodified for production calculations, and one with the dummy rows added for LandBOSSE BOS calculations.

[brtietz]

Thanks for filing. I'm closing this as "won't fix" for two reasons:

1. We haven't had funding to maintain LandBosse within SAM and keep up with more recent versions (including with things that would cause this issue) See Remove landbosse models? #1942 for more details
2. We're in the process of updating our old PythonHandler class to PyInstaller for ease of maintenance. This integration was a dependency of the PythonHandler class.

#2154 Removed this code, and given other constraints we don't anticipate having LandBosse in the 2026 version of SAM. If the funding and maintenance situation changes we would consider adding it back into a future version.