"Close" here is relative. If you look at this explanation: Average distance between two points on a sphere, along with this value from Wolfram Alpha, you will see that the expected distance along the surface between two random points on the surface of Mars is 5335 km.
But the trek between to two landing sites is 3200 km.
Was Mark Watney just lucky, or is there an explanation given for why the two landing sites are closer together than random chance?
It's not so coincidental as it might seem. There are some things to keep in mind when going to Mars:
Mars is small: Mars has a circumference of about 21,344 km; no two points can be more than 10,672 km apart. The surface area is about 144 million km^2. As you stated, expected distance along the surface between two random points on the surface of Mars is 5335 km, but the locations of the two sites are NOT going to be random, and in fact they are relatively limited.
Mars is lopsided: For some reason, the majority of lower elevations (with the exceptions of Hellas and Agyre Planitias) are in the northern hemisphere of Mars, and the majority of higher elevations (with the exceptions of some of the volcanic regions) are in the southern hemisphere. There is precious little air on Mars, so the lower you go and denser the atmosphere, and the more air you can grab with your parachute, the safer your landing. Cross off the equivalent of one hemisphere to eliminate areas that are too high. You now have 72 million km^2.
Keep warm and sunny, and go someplace easy to leave: The temperature at the equator in summer generally doesn't go much below -100 F at night, while the poles can get down to -225 F, so cold that the carbon dioxide atmosphere freezes into dry ice. The farther from the equator you go more power it takes to keep warm. In addition, the easiest (and probably lightest) way to get power on Mars is by solar panels. You want to stay places that get lots of sunlight, especially given that Mars gets much less sunlight than earth. Cross off everything more than 30 or 35 degrees from the equator. (Also, you get a little extra boost in getting back to space the closer you are to the equator. It's like free fuel.)
At this point about 62% of the surface of Mars has been eliminated from consideration. You are left with about 55 million km^2, which is close to the equivalent of the combined areas of North America and Africa, in two fairly distinct regions. 'Africa' (Amazonis and Arcadia Planitia) and 'North America' (Chryse and Acidalia Planitia) are actually connected somewhere around 35 to 40 degrees north of the equator, and you might be able to drive from one to the other without getting 'too far north', but the main masses aren't really convenient to each other.
For some reason, 'North America' seems to have gotten a little bit more than its fair share of interesting spots. It's not too big a leap to imagine that two or possibly even three of the Ares missions have landed or will land there. The possible locations for the next Ares mission are either relatively close (generally within 4000 km), or relatively far away (Gale crater, for example, is about 8500 km driving distance).
What Watney did was drive the equivalent of driving from Moose Jaw, Saskatchewan to Fayetteville, NC, which few people would consder to be 'coincidentally close'. But while a car could make the trip in 30 hours according to Google (not accounting for traffic and pit stops), the rover traveled over land at a top speed of 25 km/hour (a little over 15 mph), no more than 4 hours a day, and every 5th day off to replenish his air. He traveled at most a little under 450 km per week (roughly 275 miles).
Very late answer, I know, but I'm replying as much to clarify a couple of things for myself by articulating them (so to speak), and putting them someplace I can reference back to.
The implication (from the novel) is that he just got lucky:
I’m lucky. Thirty-two hundred km isn’t that bad. It could have been up to 10,000 km away. And because I’m on the flattest part of Mars, the first 650 kilometers is nice, smooth terrain (Yay Acidalia Planitia!) but the rest of it is nasty, rugged, crater-pocked hell.
That being said, there may have been some operational benefit in having the two sites close to each;
