# ERA5-land snow initialization

**URL:** <https://forum.ecmwf.int/t/era5-land-snow-initialization/1322>\
**Category:** C3S - Datasets and Usage\
**Tags:** ifs, c3s, era5-land\
**Created:** [10 December 2021 17:26 UTC](https://forum.ecmwf.int/t/era5-land-snow-initialization/1322 "2021-12-10T17:26:12Z")\
**Posts on this page:** 1\
**Page:** 1

<div class="post-metadata">

**Author:** ![Michael\_Shaw](https://forum.ecmwf.int/letter_avatar_proxy/v4/letter/m/8edcca/32.png) [@Michael\_Shaw](https://forum.ecmwf.int/u/Michael_Shaw)\
**Post date:** [10 December 2021 17:26 UTC](https://forum.ecmwf.int/t/era5-land-snow-initialization/1322/1 "2021-12-10T17:26:12Z")

</div>

Hello. &nbsp;Can someone describe or point to how ERA5-land gets initialized and how often (e.g. once at the very beginning of multi decade run? &nbsp;or once every day?)? &nbsp;And from what fields? &nbsp;(ERA5 analyses, e.g.?) &nbsp;And, while I understand that ERA5-land does not use any land DA (SEKF for moisture, 2DOI for snow), and that it uses ERA5 meteorology for forcing while being uncoupled from the atmospheric model, I do wonder if ERA5-land updates snow fields periodically (e.g. from ERA5) to control model drift and over-/under-accumulation through long periods; the fact that ERA5-land appears to be run on 24 hour cycles also suggests to me that perhaps this is in part to do any such updating/"initializing" for each day's fields that might be especially susceptible to such model drift issues over long periods as snow can be.

Thank you.