"We cannot continue to operate instruments and missions whose time has passed, because I won't be able to put something like InSight on Mars in 2016 ... I have to make choices."
Because the MERs landed on Mars a decade ago, the technology in them is at least a decade and a half old. Think about it: That was before the iPod, before most people had multi-CPU computers, and still well before digital cameras were mainstream. Contrast that with what one has available now to build a new mission, and you start to see some of the issues.
On the one hand, you have a still-operating mission. It's there. On Mars. Still returning usable data. The cost to build it and launch it and make sure everything is working right is done, paid for, and will never have to be paid for again.
But, on the other hand, the money to keep it going prevents new missions, with better technology, with a new and different goal, from even getting off the ground and having the potential to do new science.
It's easy to say, "Well, just increase NASA's budget!" Easy to say, not so easy to get Congress to do. Delays and cost over-runs in both the James Webb Space Telescope and the Mars Curiosity rover both were not made up by Congress and so ate into other parts of the budget, such as basic scientific research funding (what I rely on to put food on the table).
If NASA were to simply "turn off" MER Opportunity, it would not be the first still-functional mission that NASA has effectively killed. There are several in recent times, but perhaps most interesting to me is the Apollo seismic network. The Apollo astronauts on later missions installed and activated seismic stations on the lunar surface to detect moonquakes. No, there are no buildings on the moon that we needed to warn, but rather whenever there was a quake - from something impacting to tides between it and Earth and the sun - the stations would pick it up and scientists on Earth would be able to use the data to slowly build up a picture of the interior of our nearest celestial neighbor. It's a little like the way an ultrasound works, and it's earthquake data on our own planet that lets us know about the interior structure of Earth.
When NASA ran out of funding for it after a few years after the Apollo program ended, they shut down the network, despite the fact that it was still returning good, usable data, and with each new quake we learned a little more about the structure of the moon's interior. A consequence of this - besides a loss to basic scientific research - is that conspiracies about NASA finding the moon is hollow have become widespread among the astronomical pseudoscientific world. The thinking goes that they found it "rang like a bell" (which had more to do with the loosely packed surface material) and therefore either shut down the network so we wouldn't know it was hollow, or they kept the network going but in secret.
Anyway, that bit of conspiracy aside, this really is a serious issue and serious question of where our monetary priorities should be. Assuming NASA's budget is static and will not be changed to keep these long-lived missions operational, then what should a good administrator do? Should they keep the proven mission going? Or should they kill it and fund the new mission? Or, should they fund both and pull money from other parts of the budget, like education, or human spaceflight, or basic research?
I have my own opinion on this, but I'll keep it to myself. I just know I wouldn't want to be in Charley Bolden's shoes when he makes that decision and has to not only answer to Congress, but to many scientists who will see their budgets cut, yet again.