# Subseasonal-to-Seasonal (S2S) Water Demand Forecast Inputs: SPEAR Hindcasts, gridMET Weather and Drought Index, Georgetown and Round Rock, Texas (CMU KiltHub)

## Facts

- Provider: Keerat Singh, Sarah Fakhreddine (Carnegie Mellon University)
- License: CC-BY-4.0 (https://api.figshare.com/v2/articles/33065162)
- Size: 10 files
- Format: Parquet (in sampled files)
- Stored at: Hugging Face, https://huggingface.co/datasets/CMU-OSPO/Data_for_Extended-Range_Water_Demand_Forecasts_Driven_by_Subseasonal-to-Seasonal
- Access: public, readable without an account
- DOI: https://doi.org/10.1184/R1/33065162.v1
- Category: Climate & Weather
- Source: https://huggingface.co/datasets/CMU-OSPO/Data_for_Extended-Range_Water_Demand_Forecasts_Driven_by_Subseasonal-to-Seasonal
- Page: https://scigantic.com/archive/b5abb3b1-552b-5b6e-8d02-240d5d40c010
- Keywords: Climate & Weather, Parquet, Hugging Face

## About

Input data behind a study of extended-range municipal water demand forecasting driven by subseasonal-to-seasonal (S2S) climate forecasts, for the cities of Georgetown and Round Rock in Williamson County, Texas. It holds daily gridMET maximum temperature (degrees C) and precipitation (inches) for each city from 1980 to 2024; monthly hindcasts of average maximum temperature and total precipitation from the NOAA GFDL SPEAR seasonal forecast model for 1991 to 2022; the weekly Drought Severity and Coverage Index (DSCI) for Williamson County from 1980 to 2024; and annual US Census population estimates for both cities from 1990 to 2023. The utilities' water demand series are not included (available from the authors on request). The Carnegie Mellon University Open Source Programs Office converted the seven CSV files to Parquet subsets (spear_tmax, spear_precip, gt_tmax_p_1980_2024, rr_tmax_p_1980_2024, dsci_williamsoncounty, population tables) that load with the datasets library. Useful for weather-driven demand models, seasonal forecast skill checks and drought studies. License CC BY 4.0. Cite Singh (2026), Carnegie Mellon University, DOI 10.1184/R1/33065162.v1.

## Load the data

```python
from datasets import load_dataset
ds = load_dataset("CMU-OSPO/Data_for_Extended-Range_Water_Demand_Forecasts_Driven_by_Subseasonal-to-Seasonal")
print(ds)
```

## Data structure

File types in sampled files: Parquet

Example file: `dsci_williamsoncounty_19800101_20250101/train-00000-of-00001.parquet`

## Open it

- In a hosted notebook with the data mounted: https://scigantic.com/archive/b5abb3b1-552b-5b6e-8d02-240d5d40c010
- Full record as JSON (schema card included): https://api.scigantic.com/api/archives/b5abb3b1-552b-5b6e-8d02-240d5d40c010
- Over MCP: `get_schema_card` with id `b5abb3b1-552b-5b6e-8d02-240d5d40c010` at https://api.scigantic.com/mcp

## Related topics

- [Parquet datasets](https://scigantic.com/topics/parquet)
- [Climate & Weather datasets](https://scigantic.com/topics/climate-and-weather)
