Block confirmation time across clusters shouldnt be compared with bitcoin block confirmation time, as within a cluster transactions will flow quickly regardless of this time. Transferring funds across clusters might take longer, but we dont see this as a problem. Current financial system works in a similar vein and we used it as a model for ClusterCoin. Analog of a cluster would be a network of local banks (local payments are processed by local banks with a prompt transaction); across cluster transactions (served by sync blocks) correspond to SWIFT transactions.
Regarding fork risk of sync blocks it is not clear that these blocks (which contain more transactions than an average block) pose more risk than any other block. What are you basing your claims on?
From your ANN I assumed the syncing was needed to make sure the clusters all have the same blockchain data:
"each cluster will process transactions independently and then periodically synchronize transactions between geographic clusters."
That sounds like each cluster will do transactions in their specific geo-location and then sync that data with other clusters to make sure all data is available and correct at all clusters.
However now you seem to be saying that each geographical location will have its own blockchain and sync blocks are only needed if one person travels from one geographical location to another, sort of how cell phone signal gets passed from base station to base station if you have a cell phone call while in a moving car. I am not seeing where the decrease in block confirmation time will occur then. If I introduce a different altcoin into each geo-location, they will have separate block chains and their confirmation times will be small only because I have limited who can be on that blockchain by geography. If I want the transaction data from another geo-location I would then have to search numerous blockchains, not just one. Also if one cluster is taken down for some reason the transaction data stored in that cluster will be lost. In fact you have similar points of failure and loss of data all over your network of blockchains.
Clusters will not be managed by a central authority but rather by wallets themselves. Whether or not a cluster should be generated will be based on a voting by miners/pools. Comparison with wireless carriers seems farfetched.
I happened to note you compared your coin to the modern banking system, one of the most centralized institutions we have. My ccomparison to wireless carriers is apt because somewhere in your coin's code you will need to have a list of clusters mapped to their associated geographical locations in order to be able to sync someone's transaction correctly to the correct cluster as they move from one geo-location to another. This is currently how cell phone towers work. That mapping is the holy grail for every surveillance agency out in the world today. With it they can track the geographical location where a certain transaction took place. You may argue that each cluster covers a large geographical area. This may be true at first but the more clusters that are set up, the less people will be covered by each individual cluster making it easy for data mining algorithms to find what they need.