# How is the Grid = "Archived Value" data generated?

**URL:** <https://forum.ecmwf.int/t/how-is-the-grid-archived-value-data-generated/14244>\
**Category:** Datasets and Usage\
**Tags:** api\
**Created:** [9 October 2025 06:21 UTC](https://forum.ecmwf.int/t/how-is-the-grid-archived-value-data-generated/14244 "2025-10-09T06:21:34Z")\
**Posts on this page:** 1\
**Page:** 1

<div class="post-metadata">

**Author:** ![Ian\_Langmore](https://forum.ecmwf.int/letter_avatar_proxy/v4/letter/i/4491bb/32.png) [@Ian\_Langmore](https://forum.ecmwf.int/u/Ian_Langmore)\
**Post date:** [9 October 2025 06:21 UTC](https://forum.ecmwf.int/t/how-is-the-grid-archived-value-data-generated/14244/1 "2025-10-09T06:21:34Z")

</div>

The [documentation](https://confluence.ecmwf.int/display/UDOC/MARS+interpolation+with+MIR#MARSinterpolationwithMIR-Gridtogridinterpolations) (or [this diagram](https://confluence.ecmwf.int/pages/viewpage.action?pageId=123799065)) on interpolation with MIR shows output grids as starting with a SH representation, then if RESOL = AV (“Archived Value”), translation to e.g. an O320 grid as an intermediate. What is the AV grid? Is data actually stored here rather than in SH coefficients?

For example, if I download the extended range hindcasts with “grid=av”, then I get a regular grid that looks like O320 in latitude, but has the same number of longitude points as 0320 does on the equator. So it’s denser than N320.