Earlier in this thread I mentioned that I was looking for stats on the distribution of ag. production in the SU. I still haven't found ww2-era stats, but found the following from a 1970's report on Soviet agriculture (https://www.nass.usda.gov/Education_and ... iction.pdf
In the figure, "ET" means European SU (west of the Urals) and AT means Asiatic SU.
"Chernozem" refers to the "black earth" fertile regions of Ukraine, Southwest Russia (particularly northern Caucasus), and the Volga basin.
As you can see with a little arithmetic, land in Asiatic Russia had about 59% of the yield, per acre, of land in the Chernozem ET.
Relative to the non-Chernozem ET, AT land had 69% of the yield or 31% lower.
Assuming these productivity disparities held in WW2 (they're land/climate based, so seems reasonable), substituting an acre of land in the Chernozem ET for an acre west of the Urals would cost ~40% of Soviet agricultural production, all else being equal (assuming, for instance, that it takes as much manpower to sow and reap poorer land as it does richer).
If the SU had lost 1/3 of the Chernozem and non-Chernozem ET in '41 (i.e. 22% of grain production by land), then AT was producing ~49% of its grain in '42 (.38/ .78 per the graph).
The remaining 2/3 of non-Chernozem ET was producing 14% of its grain, while remaining Chernozem was producing ~37% of SU's grain.
Losing the rest of the Chernozem ET would therefore imply a ~15% ( -0.41 * .37) decline in food production, if the SU moved its agricultural manpower/capital resources to new ag. lands in AT (which is what happened during the war - SU sowed new lands in Central Asia and Siberia to make up for losses).
If the SU lost all the ET (i.e. is pushed back to Urals) then the decline in food supply - again assuming use of the same manpower/capital on AT land - would be ~19%.
The foregoing assumes that the disparity between Asiatic and European/Chernozem land in SU was equal to the disparity in 1976 when this study was published. It is entirely possible that productivity would have diverged in some respects, but it also true that during WW2 the Chernozem ET was the SU's most productive land.
My quick calculation shows a 15% decrease in food supply from substituting Asiatic land for Chernozem and 19% for substituting all European land for Asiatic (if SU is pushed back to Urals during '42).
That's in the middle range of my "shot in the dark" estimates of the consequences of further territorial losses in Southern European SU (10-25% decline in ag. productivity), so I'll give myself a pat on the back.
Of course the SU could have ameliorated the loss of good land by bringing more poor land under cultivation, but that would have required enormous manpower resources. The wartime SU already had ~55% of its workforce on the land; a 19% decrease in land productivity would have required moving another 13% of the labor force into food production ( .55 / .81 = .68 ). Holding the ratio of industry:soldiers constant (so level of armament stays constant) that would imply shrinking the army by ~29% (.13 / . 45 ).
...which would plainly have serious consequences.
[note that all of this analysis is based on a constant-sized population, in fact the total population size would have shrunk with loss of territory. This doesn't impact the proportion of people/workers/soldiers for a given per-capita level of arms/food, but it would impact the size of the army dramatically as well].
Now I'm not claiming that these figures are exactly right, but if they're even ballpark they demonstrate why the SU could not afford to lose much more of its European territory without losing its ability to feed/arm its populace and forces.