In extreme long range shooting consistent velocity is a vital component of hit probability. To hammer this point home consider this: The radius of a standard 36" ELR target is only 18 inches. Put another way, any increase or decrease of velocity that would create 18" greater or lesser of drop would result in a clean miss, even if the shot was performed perfectly. In the case of this 33XC (300gr Berger OTM @ 3,060fps) at a target distance of 2,500 yards, that velocity change is less than 8fps!!!
Obviously, one component that must be considered during load development is the extreme spread and standard deviation of loads. In the case of the 2,500 yard target engagement, an ES of 15 would equal the diameter or height of a standard 36" ELR plate. Anyone who has load developed an ELR rifle or put one across a chronograph, knows that such an ES would be rare for long shot strings like those fired in an ELR competition.
An SD of a load with such an excellent ES would likely be below 6 feet/second. By using Applied Ballistics' Analytics software this would produce a hit probability 79% if all other variables are accounted for, the rifle is capable of shooting in the ones(<0.2 MOA), and the shot is executed perfectly. These variables include, but are not limited to, wind call, atmospherics, Coriolis, and powder temperature!
Shooters typically calculate their ES/SD at one temperature. If the effect of powder temperature is not factored into a shooting solution these numbers essentially get a lot larger. If powder temp changes the velocity by 20 fps then the ES is changed by at least that much. If this change in velocity is not accounted for, it would essentially double the SD, if not worse, bringing the hit probability down below 50 percent. A more precise way to view this error would be in target subtension. The target (36"^2 @ 2,500yds) subtends at 0.40 milrad and 20 fps at this distance equals 0.48 milrad.
Aiming center target and not accounting for the muzzle velocity increase due to powder temperature would result an average point of impact shift over one-half target off of the target's edge. Since the vertical dispersion is equal to more than the full height or diameter of the target, another way to imagine this error's effect would be to reduce the target's height by at least half or, depending how the ES/SD shakes out, into fourths which would bring the hit probability down below 25%. Oddly enough, in this scenario, the better the SD, the worse the hit probability would get. It might be easier to show in graphic form at what distances do increasing SD's become detrimental to hit probability.
Hopefully the scenarios above are adequate in convincing shooters to do their best to quantify and account for powder temperature sensitivity. As most shooters know, different powders exhibit varying degrees of powder sensitivity. The powders most notable for being insensitive to temp swings are the Hodgdon Extreme powders. The only powder in this line suitable for the 33XC was H50 BMG. Unfortunately, this powder is going to get increasingly hard to find. After extensive pressure testing and load development the Alliant Powder Reloader 50 was chosen. This powder produced the best groups and also the highest and most consistent velocities without creating unsafe pressures.
The next step was to test RL50 for temperature sensitivity. The procedure for this was relatively straight forward. Five groups of five rounds of ammunition were brought to temperatures from 45°F to 110°F. These were fired across a MagnetoSpeed chronograph and at calibrated targets. The Strelok Pro screen capture below shows the results:
Compared to H4350 it could be said that RL50 is temp sensitive. This statement would be unfair for many reasons. H4350 would not be used in the high volume cases that RL50 would be and this may affect burn characteristics. H4350 exhibits temp sensitivity factor of around 0.2% in the common 6.5mm cartridges like the Creedmoor and 260 Rem.
Finally, to check the accuracy of this test another 5 round group was shot at the ambient temperature present at the end of the test. According to the Applied Ballistics Kestrel Elite used for the test, that temperature was 82°F. By using both the Kestrel's temp table and Strelok's, the estimated velocity was between 3,054fps and 3,057fps. The five rounds averaged 3,056fps! This test was repeated at 100°F and both programs predicted the average velocity within 3-4fps!
It should also be noted that SD's of a particular load at one temperature are not likely to be consistent across all temperatures. This load exhibited better SD's at temperatures near the mean tested and worse at the extremes.