Air Temperature 2017 at 2m India

About Dataset

This dataset provides 3-hourly forecasts of air temperature at 2 meters above ground level, for the Indian region during the year 2017. The data is generated using the Global Forecast System (GFS) T1534 model, with a high spatial resolution (~0.125°), archived by IITM Pune under the MoES. It contains 10-day forecasts with 81 time steps.

How to Access the Dataset:

1. Review the Dataset Info Page
On redirection, you’ll land on a URL similar to:
Sample example: https://ardc.tropmet.res.in/thredds/dodsC/data/v10/gdas_2000_india_v10.nc.html
This page provides:

• Metadata: file size, variable names, and dataset conventions.
• Interactive Access Form: inputs for subsetting and extraction.

2. Locate the Variable of Interest
Scroll to the variable (e.g., VGRD_10maboveground) listed as a “Grid”. This is the dataset’s main content (e.g., wind speed, temperature, etc.). Check the box next to it.

3. Fill in the Data Subsetting Parameters
You must manually fill in five fields to subset the data grid:

• Time: Index of forecast issue date in the dataset. Use 1:1:2 to extract data from Jan 1 to Jan 2.
• Forecast_time: Indexes the forecast horizon. Use 0:1:80 to access all 3-hourly steps for 10 days.
• Latitude & longitude: Leave blank to extract full India coverage (default ~8°N–38°N, 68°E–98°E). You can enter ranges to subset, like 50:1:100 to access a slice.
• Variable selection: Ensure only one variable per card is selected to prevent data overload and errors.

4. Data Format Selection
At the top of the page, select:

• Get ASCII for plain text viewable in-browser or scripts.
• Get Binary for machine-readable scientific applications.

These generate the requested data instantly in the selected format.

5. Download the Output
After clicking “Get ASCII” or “Get Binary,” a new page or download will start. You can copy or save the returned raw data to your machine or pass the OPeNDAP link into software like Python (using xarray or netCDF4) for programmatic access.

Important Notes:

• Time and forecast_time are index-based (not real dates). Time = 1 refers to Jan 1, Time = 2 to Jan 2, and so on.
• Forecast_time ranges from 0 to 80 representing 3-hour intervals up to 10 days.
• This dataset contains a single forecast file per day, so only one “time” index should typically be active per request.
• Avoid selecting multiple variables on this form — each dataset card is designed for a single-indicator download.