This suggests that a large fraction of 2009 coins are not being hoarded, but have been lost.
What is it that leads you to that conclusion?
the fact that the proportion of unspent coins is higher for 2009.
I don't see why coins from 2009 would be hoarded more than coins from 2010. Because some of the innovators (like Satoshi (???BTC) and mbknight (371k BTC) have hoarded there early mined bitcoins and never used the majority of them to date.
So, coins were hoarded because early users hoarded them? This sounds like a circular argument...
Many people fantasize about Satoshi hoarding coins. Here is my interpretation of the data:
During the whole year 2009, difficulty was equal to 1 (the minimum). There was not enough hashing power to make difficulty increase, and the number of coins mined was, on average, lower than 1 block per 10 minutes. Now if you look at the graph, you'll see two episodes in august 2009 where the hashing power decreased dramatically, almost to zero. These episodes are short, but they are clearly visible, and we can see them both on the difficulty graph (bitcoin.sipa.be) and on my velocity graph. What this means is that most users stopped mining during these episodes, either because they lost interest, or because they were not able to mine for some other reason.
I believe that Satoshi was among those who continued to mine during these episodes. This hypothesis is based on the fact that he is the creator of the network, therefore he was the less likely to "lose interest", and he was also the most likely to do everything to keep the network alive. Think about it: if you created Bitcoin, you would probably consider it as your "baby", and you would keep caring for it even if other people decided to leave. The fact that Satoshi was still active in 2010 confirms this.
If you accept that hypothesis, then the blockchain graph shows two things: First, it gives you an upper bound on Satoshi's mining power in 2009; it was not greater than the height at those dips (it may have been lower than that, if he was not the only person who kept mining during those dips). Nevertheless, the fraction of 2009 coins that could have been mined with that hashing power is very small, less than 5% of the total 2009 output (naked eye estimate).
Second, the proportion of 2009 coins that have been moved from their initial address corresponds, more or less, to the fraction of mining power remaining during the dips (height of the green curve at those two dips). This is consistent with the idea that the people who mined during the two 2009 dips did believe in Bitcoin's future, and thus they did not lose or delete the coins they mined throughout 2009.