Feb 28, 2015

New Product - ESRL Style Maps

I have always wanted to produce those fancy weather maps that I see all over the Internet. I read some how-to but never fully invested. I am lazy. Instead of producing them myself I figured I would outsource. The ESRL provides the perfect setup. Utilizing some of my new tracking output I have introduced ESRL style forecast maps.



I am currently producing 500mb Geopotential Height anomaly, Surface Temperature anomaly, Surface Sea Level Pressure anomaly, and Surface Precipitable Water anomaly maps. Forecast ranges include 6-10, 11-15, 16-20 and 21-25 day. The forecasts can be found here. For the fun of it, below are the most recent (maybe) daily temperature anomaly forecast maps. Click here for more.

USA- 6-10 Day Forecast - Surface Temperature


USA- 11-15 Day Forecast - Surface Temperature


USA- 16-20 Day Forecast - Surface Temperature


USA- 21-25 Day Forecast - Surface Temperature


Like all of this junk, it is a work in progress. If there are any questions, comments, or suggestions on the material presented please let me know. Thanks for reading!

Feb 23, 2015

A Transient Cycle Length with Multiple Phasing Cycles

So where was I? Ah yes, I dropped the high amplitude and frequency states (ISO short-term component, 10-20 days) from the CONUS output and plotted the 30-90 day correlations exclusively. More

The map below shows the RAOB stations that I collect data from. I don't collect Canada or Mexico and it seems I randomly do not collect in the deep south US. I blame it on laziness. I am super lazy.



My thoughts are not well organized or educated. They are likely difficult to follow. I will attempt to explain them in short detail. The image below shows which region the highest correlations stem from. The range is 0-1. The higher the value, the more high correlations stem from that region. Example; the daily analysis for 2/21 shows region 5 has 9% of it's stations reporting a top 10 value. Region 8 60% and region 9 67% of it's stations in the top 10. This correlation could be considered "east based", where this region shows the most correlation.



These correlations are charted in a heat-map like table form. The image below shows the first 21 days of February. The left most table is the top 10 cycle lengths, listed from left to right, 1 through 10. The table directly to the right is the corresponding correlation values. The 3 tables to the right are mode, median, and average of cycle lengths for the previous 30 days, since December 1st, and since August 1st. The entire table can be found here.



A quick analysis of the heat-maps suggest a transient cycle length with multiple phasing cycles taking place. Similar to a standing wave. If there are any questions, comments, or suggestions on the material presented please let me know. Thanks for reading!

Framework: Use current NOAA/ESRL Radiosonde Database to analyze large-scale upper atmosphere patterns in standing wave notation. Described specifically to harmonics, reflecting the temporal/transient behavior of the frequency wavelengths in correlation and relating Intraseasonal Oscillation to Mid-Latitude recurring weather patterns.

Goal: Forecasting skill of upper-air and surface weather trends.

Doug Heady Joins Gary Lezak at Weather2020

Great to read that Heady and Lezak are back together. Gary puts together a wicked map comparison in a recent blog. Can someone point out 4 similar features in the maps just 9 days prior to this latest map comparison? I have provided the map analogs below.





I find it fascinating what Gary and Doug see. If there are any questions, comments, or suggestions on the material presented please let me know. Thanks for reading!

Feb 13, 2015

I dropped the high amplitude and high frequency states and plotted the 30-90 day correlations exclusively

I dropped the high amplitude and frequency states (ISO short-term component, 10-20 days) from the CONUS output and plotted the 30-90 day correlations exclusively.



What is this junk? Back when I followed Lezak's Recurring Cycle and/or the Heady Pattern, comparing 500mb maps with the naked eye was the only way to find/follow the cycle. I quickly learned everyone has a different eye. Instead of looking for blessings from the cyclic masters I decided to seek the cyclic nature in data. Much like how they (LRC/HP) compare maps I decided to compare legitimate peer reviewed oscillations in a similar manner. Rather than look for a recurring cycle at ~50 days every couple days (or what is perceived as the cycle length) in maps, I formulate correlation of a 10 to 90 day range from sounding data on a daily basis. I find the dominant recurring patterns and plot the output.

A correlation example. Say I am finding the correlation for the date 9/1, for day 20 in the 10-90 day range. The data arrays that are used to find the correlation consist of data within the dates 9/1 thru 8/12 and 8/11 thru 7/22. I find the correlation for day 10 through day 90 and plot the highest correlations. This is what is plotted in the tables of the image above. The left table is the region* the top correlations in the middle table originated from, the middle table is the top 10 oscillation frequencies (aka "cycle" length) and the right table is the corresponding correlation value of the middle table. At the very bottom of the table is mode, mean, and average statistics as well.

*I categorize the CONUS RAOB stations in regions of climate. I simply label them region 1 through 9, see image below.


Framework: Use current NOAA/ESRL Radiosonde Database to analyze large-scale upper atmosphere patterns in standing wave notation. Described specifically to harmonics, reflecting the temporal/transient behavior of the frequency wavelengths in correlation and relating Intraseasonal Oscillation to Mid-Latitude recurring weather patterns.

Goal: Forecasting skill of upper-air and surface weather trends.

This is just quick summary that stemmed from this twitter thread. I will likely add to the entry in the future. If there are any questions, comments, or suggestions on the material presented please let me know. Thanks for reading!