About HREF

What is the HREF

The High Resolution Ensemble Forecast (HREF) is an ensemble of several different weather models running at ~3 km horizontal grid spacing. Instead of making a single forecast Ensemble forecasting uses several different forecasts which aims to give an indication of the range of possible future states of the atmosphere.

What am I looking at?

Accumulated liquid equivalent precipition (rain + if you melted all the snow) from the HREF.

Why?

I feel this ensemble gives a better view of what amount of precipitation to expect in mountainous terrain, but nobody was making plots for this model I found acceptable.

How do I convert precipitation to snow?

For intercontinental areas like Colorado and Utah I find multiplying by 15 translates well to snow report. So 1in precipitation = 15in snow on the ground. Actual ratios can vary greatly.

36 hours isn't enough to plan my powder days.

NOAA is planning to increase the forecast length to 48 hours in Feburary 2021.

What is 00z and 12z

The initialization time of the model in UTC time. 00z is midnight UTC

What is sprd and lpmm.

sprd is the ensemble spread. NOAA does not specify more than that. I think might be the difference between the highest and lowest ensemble members

Research has shown the PMM/LPMM are more accurate than the mean. According to NOAA SPC HREF Viewer Website - Conceptually, PMM (probability-matched mean) is a variation of the ensemble mean with the original ensemble amplitude restored. At each grid point, the ensemble mean value is replaced with a value from the full distribution of individual member forecasts whose rank matches the point's rank within the ensemble mean distribution. See Ebert (2001, MWR) for more details. LPMM (localized PMM) is a new technique from Clark (2017, WAF) wherein the PMM calculation is restricted to grid points inside some radius of influence, preventing precipitation in geographically distant areas from influencing the local value. On this site, we use r=110 km.

About Danwrf

What is the Danwrf

Danwrf is an instance of the Weather Research and Forecasting (WRF) model I am running on a spare computer.

Do you know what you're doing?

I think I have a good enough grasp to create a reasonable model configuration. NOAA has done the hard work of assimilating weather observations across the globe to create accurate snapshots of the atmosphere and running a global model which I use for initial conditions/boundary conditions which enables me to to run the model.

How does it compare to NOAA models

NOAA uses the WRF in many of their models including all but one member of the HREF. NOAA has to ensure their models work well for safety/economics reasons so are conservative with model configuration and updating WRF versions. I try to use reasonable configurations but will experiment with some cutting edge options and try to use the most recent WRF version.

If NOAA is already running a bunch of WRF models why are you running your own

Mostly personal interest. Also NOAA has to provide forecasts for the entire country so their limited resources are spread across forecasting for the entire United States. I only care about where me and my friends ski so I concentrate all my computer resources to forecasting in those areas.

What is the difference between the Danwrf 15km and Danwrf 5km

The WRF divides the forecast area into a grid where each square has one forecast. The 15km is a model where each square is 15km x 15km. The 5km has 5km x 5km squares and runs within the 15km model. At 15km resolution the topology of mountains is not captured well at all and the mountains look like smooth blobs of higher elevation. The 5km resolution starts to show distinct ranges but does not capture things like narrow vallyes. The mountains look like a smoothed wall, but is much better than 15km resolution. The nesting of the models allow me to run a more computationally demanding higher resolution model with limited computing resources.

About MPAS

What is the MPAS

Model Prediction Across Scales (MPAS) is an Earth modeling system primarly developed by NCAR and Los Alamos National Lab. The defining feature is is MPAS uses unstructured Voronoi meshes for model cells which you can think of as variable sized hexagons. Other models like the WRF and GFS use a rectangular grid. The hexagons fit better on a sphere compared to rectangular cells

tldr; It's a new weather model with some cool properties