When you look at graphs of new daily infections on the national level, what you see is not the entire picture. For instance, here is the most recent MAPS forecast for the next four months. It appears to be a fairly smooth curve. However that is not the entire picture. Let's look a little closer at the state level. I seperate them by population size. Those >6 million and those less than 5 million. On the left is the larger states. What should be clear is that they all have different temporal distributions in terms of onset and peak. Now let's look at the

MAPS USA Ensemble (N=20) Forecast 1-6-2025 to 5-6-2025 with Masking Results

This is the first forecast done with a newly added compartment that is used for immunity profiles from natural immunity with/without prior vaccination. We will see how it works out, but historical runs indicate better agreement that the prior version of the model. The figure below shows the +/- 95% CIs (thick black lines), the ensemble mean (thick red line) and other ensemble members within the bounds. If this verifies it would be an unprecedented early spring wave. This is a

We have patched the WHN wastewater-derived infections for the USA, and things look much better. There were six wastewater stations in one county of NY that were a couple of orders of magnitude too large. It appears we peaked on January 10th, 2025 at ~1.31 million new daily infections. I can now run the model using the last good date before preliminary estimates are employed (1.15.2025). In addition I will be able to include the latest variant proportions update. Will this be the last time I am able to do a forecast after yesterdays bad news?

Here is the first of many MAPS forecasts over the coming weeks. I am working on some things to improve the model. If this is verified over the next 4 months, it will be unprecedented. The forecast beyond Mid-February is subject to considerable change depending on emerging and is characterized by the increased spread in the forecast. The two emerging variants employed by MAPS are XEC.4 and LP.8.* based on sparse data. I attempt to account for the estimates by varying their traits through the integrations as well as their temporal appearance.

Here is the most up-to-date wastewater data we "trust" (). We are inclined to believe it lies closer to the maximum calculation (see webpage). The right-hand panel shows what MAPS had forecast three weeks ago. I am working on a new forecast with today's whn.global/estimation-of-…
new variant proportions data. Should be completed this weekend. I remember about 2 months ago surmising that due to the massive summer wave we would see a well below normal holiday wave. Still when MAPS indicated minimal Thanksgiving bump I was concerned something was wrong

@threadreaderapp unroll Sshh, snuck away from my ICU vigil. Justin, this study home with my PhD in atmospheric science and ecology😊 some thought I was silly putting weather in my pandemic model. I’ve done several studies with no met, 50km resolution monthly climatology, ~12km

https://twitter.com/ejustin46/status/1788060572610634066

Daily forcing and compared 2.5 years of simulation from 3-1-2020 to 9-1-2022. Ensembles (n=20) integrated on a 1km grid covering North America. The 12km met forcing was clearly superior, followed by no met, and last was the climatology ( yes it makes sense). Pandemics are global