Simply put, the Soviets advanced too quickly. In two weeks they'd gained an area the size of Western Europe, and their supplies just couldn't keep up, even then the advance only stopped when supply from the air became unfeasible. An invasion of Honshu would have different concerns, namely the limited capacity of the Pacific Red Banner Fleet, rather than the rate of advance.
The deteriorating food situation will likely lead to some uprisings, (the Japanese predicted famine by early 1946) and a Communist revolution was a serious concern for the Japanese government. That said the insurrections would likely be localised and more malnutrition based rather than Marx, however the Soviets would certainly try and spin it as a declaration of support for the latter.
Soviet invasion of Japan's an interesting topic, I've got stuck until dawn reading threads of the topic on here including some you've commented on The Red. Ended up reading about it through the books people had mentioned on Amazon preview till dawn ha.
Here's a photo of Rumoi, the landing site for the invasion -
http://commons.wikimedia.org/wiki/File:Rumoi_city_from_the_sky.jpg
This page mentions orders as to initially securing only the Island North of Rumoi (
http://prokarelia.net/en/?x=artikkeli&article_id=600&author=62)
It's a nice little village (not a city till 1947), has a sheltered harbour. There are two roads from it. One hugs along the coastline North-South where Cruisers/Destroyers can intercept opposition.
The other leads inland 50km upriver to Fukugawa where it bisects the main road links between two of the three major Japanese positions at Sapporo (HQ) & Wakkani (other defenders are far over at Kushiro on the East coast), (across the straight from Sahakalin. Of note is the 'B' class - most of them were 'B' coastal defense fortress at Wakkani for the regular seeming comments in other threads about it being the logical place to invade. Pair of 11" guns in a former battleships armoured turret removed after London arms treaty limitations just to start with
) which are much further away and.. bisects the shortest rail link (the importance of which should become apparent below).
http://en.wikipedia.org/wiki/File:JR_Rumoi_Main_Line_linemap.svg
It also seems to be just within the combat radius of IL-2's operating off Sahkalin.
I can start to see the thinking behind it.
Projections from ammunition supply, & the rate of casualties encountered elsewhere left one commenter on a thread here thinking the Soviets would conquer Hokkaido within 2-3 weeks.
This map in one thread was based off proposals for the division of Japan going around in American circles before the bomb was dropped.
http://upload.wikimedia.org/wikiped...x-Proposed_postwar_Japan_occupation_zones.png
The more & more I look at it, a 'People's Democratic Republic of North Japan' following an Imperial Japan that doesn't surrender after the Bomb looks more & more plausible.
___
As to Communism & food.
The Strategic Bombing survey later noted that the near famines of the post-45 period could have been far worse. What grain was there needed to be transported by rail. 97% of the highways were unpaved (note above, for an invasion of Hokkaido with logistical considerations, rail seems the only effective way to move troops. Given projections from preparations for American invasion on either Honshu or Kyushu, marching seems the only other way, for those locations it was expected to take 60-90 days. Equivalent marching distances over Hokkaido look like they take over a week just to meet the beach head. Sounds like it's Home Guard vs dug in Marines/Paratroopers, maybe tanks).
The Japanese coastal transports were mostly sunk and the rest would have been sunk within weeks. There was no redundancy in the rail network, on Honshu it often ran just along the coast. There was no major planning for repairs or available materials for them, even wooden rail sleepers would have been a problem. The barges that delivered coal from Hokkaido to Honshu to run the trains were hit by Hasley's task force, were down to 18% of their original capacity. The Americans would have hit major rail links on Honshu in August and paralyzed the system. There was a four day stockpile of food.
OTL major American aid prevented starvation in Japan by Jan-Feb 1946 with it's (Japan's) collapsed harvest. A longer war with the destruction of the railways might prevent it's effective distribution even with the best of intentions (most famines result not from there being not enough food but from it not being in the right place). The shortfall would have been on the order of 40% even with starvation rations in effect. This in a country of around 100 million people.
Note domestic resistance has been crushed - the Communist agitators/ringleaders - i.e those who've actually had a go at reading Marx etc & have any kind of sophisticated understanding and organization, not the average person with Communal/Utopian notions you get in rebellions - have been rounded up & shot in the 20's by the Army. There's likely going to be rebellion as soon as the powers that be can't feed the people, but it might resemble peasant rebellions of the past or openly warring factions of the leadership. The relevant term the Americans had for the scenario was Japan 'Unraveling' rather than conducting an 'Organized Surrender'.
Some hints of American plans involved shifting the Invasion to Northern Honshu once preparations were discovered on Southern Kyushu & the Kanto Plain (+ Soviet advances).
Whether the U.S land in Kyushu or Kanto or Northern Honshu, the key seems the timing of Olympic. If the coup plotters can keep the people from rioting for food in Sept-Oct until the Typhoon/U.S landings then I think they can use the excuse of the 'divine intervention/emergency' to keep their rule going until some time in 1946. Olympic being initiated makes a Soviet invasion of Northern Honshu the following day seem likely. If they can take Hokkaido in 2-3 weeks that's about 6-7 they have to prepare for it if not get on with it.