Hello everyone,
Can observational nudging be used for real-time forecasting?
So far, it does indeed improve nicely my forecasts with respect to the measurements, but my simulations are for in the past, with the observational data ingested being already available. I am using GFS forecasting data.
Now I am running simulations with obs_idynin for dynamic initialization, to turn on the ramping down function that switches off the nudging before the actual forecast (from the WRF v4.4 User Guide Chapter 5, page 108). But when I do that, the simulation with nudging starts to coincide with the simulation without nudging very soon after the ramping down is finished.
So how does a typical real-time forecasting system with nudging look like (just as a stream of thought), given that the observations are accessible live?
Thank you in advance!
Can observational nudging be used for real-time forecasting?
So far, it does indeed improve nicely my forecasts with respect to the measurements, but my simulations are for in the past, with the observational data ingested being already available. I am using GFS forecasting data.
Now I am running simulations with obs_idynin for dynamic initialization, to turn on the ramping down function that switches off the nudging before the actual forecast (from the WRF v4.4 User Guide Chapter 5, page 108). But when I do that, the simulation with nudging starts to coincide with the simulation without nudging very soon after the ramping down is finished.
So how does a typical real-time forecasting system with nudging look like (just as a stream of thought), given that the observations are accessible live?
Thank you in advance!
Last edited: