Dwell time is a concept used heavily in the somewhat specialized science of targeting and tracking sensors. The concept itself is quite simple; the rudimentary definition is "the time in which the beam of an active sensor remains focused on a single target." The actual relevance isn't quite as obvious.
Dwell time is significant because it directly affects both the limits of the sensor's ability to track targets, as well as the ability of the sensor system's ability to multitask. Let's take an example. Let's use a fictional air defense radar for the moment; we'll call it BUCKLER. BUCKLER has two jobs; the first job is to look for incoming targets such as missiles or aircraft (search radar), and tell its users about them. The second job is to track targets that it has found, so that its users can both glean additional data about the targets as well as (if they decide) fire guided weapons at them.
Okay. BUCKLER, for the purpose of this exercise, is mounted on a ship, and looks (for simplicity's sake) at only one 'arc' - the forward 90-degree facing arc, centered on directly ahead. In other words, it can 'see' targets that are no more than 45 degrees to the left or right of the ship's nose. Our first question is: how far out will BUCKLER be able to see targets?
Now, this is a much more complex question to answer fully than we can tackle here, but let's do some back of the envelope calculations. The answer to this question lies in two very important numbers; the first number is the minimum radar signature that BUCKLER can detect. The second is the strength of BUCKLER's radar beam. There is a third number that we'll need, namely, the radar reflectivity of the target; however, typically, for questions such as this, we'll be given that information. So, roughly, we need to determine how close a target with the given reflectivity can get to BUCKLER before BUCKLER's radar beam, bouncing off of it, is reflected back to the ship with a strength high enough for the receivers to detect it. Of course, we could simply do a square-law calculation based on the signal strength; however, that would be
insufficient! While that might tell us how far away we could see a target that was approaching BUCKLER with the beam aimed straight at it, that would assume that BUCKLER spends all its time watching the same point in the sky - coincidentally, the same point the target arrives at. BZZZT, nope, not true. This is, after all, a search radar, at least some of the time.
Okay, so now what? Let's take a quick look at how a radar system works. It doesn't just radiate signal in all directions; that would be wasteful! Rather, it spends a limited amount of time with the beam pointing in a particular direction. It 'sweeps' that beam through its search field - 90 degrees in our case, remember? - over and over. Let's
make this simple, and assume that BUCKLER's radar is a strictly horizontal scan. In other words, at each moment, the radar can 'see' a narrow slice of sky from top to bottom.
Perhaps the radar is built so that it can be moved across the target arc in 1-degree increments.
So, at any moment, BUCKLER might be in one of ninety different positions, looking at one of ninety different one-degree-wide slices of sky. For completeness sake, when in search mode, BUCKLER scans back and forth constantly, looking at each degree for a small piece of time. Here is (finally!) where the concept of dwell time sets in. Let's say that the system is set up so that if it receives 0.01 watt of radar energy at the receiver, there may be a target in that direction. I won't go into the calcs to get that number; it's pulled straight out of...well, never mind. :-) However, remember that any antenna will always be receiving some form of signal, if from no other source than the sky (the sun is a great source of microwaves!). So in order to avoid fooling yourself, your scientists and engineers have determined that in order to adequately distinguish itself from the noise, a target must reflect back 0.01 watt to the antenna. For more information on how far out it can see things, see Palpz's excellent Radar Range Equation.
So we scan the sky. Out of our 90 1-degree sweeps, we get four returns (let's say). Now comes the hard part - verifying those signals. We could, if we wanted, just continue to do sweeps and then keep track of the returns, to see if they continue to show up. However, at some point, we'll want to figure out more information about them, like their range, speed, direction, maybe even their shape. Also, what if they're moving across our view? Keeping track of them will get complicated.
In order to do this, we'll have to spend more time 'looking' at them than their 1-degree slice would normally get. This is because in order to get more information, we need more
signal! We need not only a certain signal level, but we need a long enough pulse to determine the Doppler shift of the signal (for velocity measurement) and we need to stick around long enough for the radar energy to go out and come back, so that we can time it and see how far away the target is.
Thus, we need to increase our dwell time.
The longer our radar 'looks' specifically at this one target, the better we can see it. If we look at it long enough, we can accumulate enough information to perhaps even identify the type of object it is by its shape! We can get a good enough sense of its direction that we can fire missiles at it, telling the missile's guidance system in advance where the target is and what it's doing.
There's a problem, however. In order to look at this target for more than our normal 1-degree sweep time, we must now shortchange some other areas of sky - or, we
need to spend less time looking at each of the other slices. So, for each arc, we're sending out less radar energy per 'look.' This means that any targets that might happen
to approach from those other arcs may not be seen until it's much closer, because the lower signal won't return 0.01 watts until the target is much much nearer. Or, perhaps, we just don't have time to look at everything, and while we're measuring the length of target 1's body parts, BUCKLER has simply stopped looking at anything outside a
20-degree arc from that target. Suddenly, our ship looks a lot more vulnerable to sneak attack.
So, (to finally conclude), dwell time is a very critical concept. Since there's usually not much you can do to change the signal strength of an active radar (outside a fairly
narrow, preset range) the only way to increase or decrease the amount of return you get off a target is to increase or decrease the dwell time - which, in turn, has a significant
effect on how far out and/or how much of an arc you can watch, and/or how many specific targets you can follow at once.
Note: This is a completely notional example. 'Real' radar systems and users have all manner of tricks they can play to mitigate the problems described above. For example, separating the radar receiver from the transmitter array helps; if you have a receiver array that is sensitive enough to watch all of your arc at once, you don't have to worry about holding position long enough to get a return. However, you'll still have to paint the target long enough to get a return with enough energy to figure out more stuff about it, and eventually, your ability to service targets and sky will be degraded. And so it goes.
The term can also be used to describe the amount of time an orbiting satellite can see or communicate with a particular point on the earth's surface. If a satellite is not in a geostationary orbit, it may have a limited 'window' where it can see a particular point.