I've argued in posts on the Drake Equation (really more a framework for trying to work out the odds of finding extraterrestrial life), that the L factor, representing the amount of time for which an intelligent civilization is detectable on a planet, is both underappreciated and overestimated. That is, it's not just important whether a planet can develop life -- which is were a lot of attention is -- but just how long a planet with intelligent life is detectable as such. If that time span is not very long, then there might well be intelligent civilizations out there that we don't know about, or have much hope of ever knowing about.
Radio transmissions are often used as a proxy for intelligent life. Clearly, if we detect a radio signal coming from planet X with a structure we can't explain by natural means, we have to seriously consider the possibility that some intelligent life form sent the signal. Artificial-looking radio signals strongly imply intelligent life, but lack of them doesn't imply lack of intelligent life.
Radio isn't the only way to go looking for intelligent life. We're already able to get some idea of the atmospheres of exoplanets based on their effect on light from the parent star as they transit between us and that star, provided everything is in a favorable alignment. That ability is liable to improve over time, to the point where we'll be able to detect whether a planet has a chemical composition that's likely to be produced by something like life as we know it. That's pretty impressive, if you think about what it entails, and we're just getting started. Astronomy has gotten really good at gleaning ridiculously faint signals from vast fields of noise, and while there are some fundamental limits to what we can gather, there's clearly a lot more we can do within those limits.
Likewise, if we can detect some signal related to a planet's surface, and we can observe the same planet from different angles (which is often possible since planets rotate) we can get some idea of any changes in the surface as seen from different angles. Similar techniques were used to get a very rough map of Pluto's surface prior to the New Horizons mission. It may also be possible to detect the polarization of light coming from a planet, and there are probably other sources of data. Put together enough such hints and we may be able to measure whether a planet has anomalously dull or shiny or hot or cold regions or similar that might indicate ... something, maybe enough to say that there's probably a civilization something like ours on a given planet, and not just an odd configuration of protoplanetary dust.
So suppose that some twin Earth in the general vicinity -- say a few dozen or a few hundred light-years -- develops along similar lines to ours. Suppose that some time after a species like ours arises it discovers radio, but not long after that it finds more efficient ways of communicating than blasting radio waves in all directions (including a tiny fraction headed towards us). In my previous posts I argued that that's probably about it. We won't be able to detect any signs of intelligent life (life, yes, civilization no) except for a tiny portion of the planet's existence, and so the odds are very low that we happen to be listening at the same time they're broadcasting.
But surely twin Earth will have cities and other large artifacts for longer than it has detectable radio emissions, and have them both before and after their radio era. Our ability to detect such artifacts will only improve over time. Suppose our techniques get to the point where we could detect the analog of a city of, say, 100,000 people by way of its structures and overall impact on the surrounding environment. There are thousands of those on Earth now and, more to the point, there have been for quite a while. Thousands of years, versus decades for radio transmission. It's at least possible that there will continue to be cities for thousands of years more. This is a couple of orders of magnitude longer than our detectable radio era might end up being, not something easy to write off.
Nonetheless, I don't know that it changes the picture much. On the one hand, even this larger time window is still pretty small on a planetary time scale. On the other hand, it's not at all clear to me that detecting a signature consistent with cities means that there are cities there. I'd want to see a lot of work to rule out natural formations that we haven't thought of, and even then city-like collections of life don't necessarily mean intelligent civilization. We are not the only life forms on Earth that can gather in numbers or have a significant environmental effect.
It's also entirely possible that we'd see the same effect with cities on Earth as with radio, just on a slower scale. That is, we or our counterparts might have cities for a long time, but not have detectable cities for very long. I'm not going to predict that humanity will necessarily lessen its overall impact on the environment over time, but it's possible. If we become cleaner and (much the same thing) more efficient, we become harder to spot, and likewise for a hypothetical alien civilization.
Nonetheless, it seems dangerous to assume that whatever impact we do have, or an alien civilization has, will be undetectable from interstellar distances. It will probably be detectable as an overall signature. The question is what could we make of such a signature. We'd probably be able to associate it with life, but what kind of life?
[ Re-reading an earlier post I see I already took this point into account, although in a more abstract way -- D.H.]