And thus has the decision made for them. It would, perhaps, have had a chance at its own evacuation, but when some of mages lose their links with the fleeing helicopters, they change targets. It goes down with all hands, ice suddenly building up even more rapidly than before.

It's not, objectively, the worst thing that happened today. The equipment is expensive, but civilization can afford to replace it, and if the true deaths are unfathomably high for normal operations they're still not the majority of those who perished in the last 12 hours. It still hits really hard.

They knew going into this that the aliens had bizarre anomalous capabilities at odds with their normal combat performance and were not in all ways best modeled as a primitive army of dath ilan's past. The ability to open portals wasn’t even an isolated example, but this was still not... what they expected to happen, and it's still a bizarre combination of incredibly devastating and incredibly restrained. The lesson is learned, though; the aliens are either bizarrely capable of downing flying machines, or else consider them enough of an escalation to escalate back, and in either case dath ilan doesn't want to lose more helicopters and crews trying. The decision to keep the fighter jets far from the city earlier is feeling very justified right now. 

The order goes out to redirect the advancing troops towards the landing zones for the parachuters, as well as encircle the crashed helicopters; it's vitally important that they advance until they can cover emergency services medical officers getting at the bodies, and hold it long enough that they can do their work before withdrawing. The assault itself is being called off, though, due to the need to reorganize and reassess. They're going to have to figure out if there's a form of air support they can and should risk, or else how to handle things with just grounded assets, and also buy time to see if they can start predicting some of these twists in advance rather than after people get killed. And they're probably, much as they dislike it, going to have to start holding back less; their earlier use of gunfire on enemy air support didn't provoke an escalation, and it's increasingly uncertain that they can win this fight without either suffering or inflicting significant casualties.

Victory! That'll teach them not to mess with the empire.

Some of the troops had been getting nervous hearing about the barbarian advances through rumor and garbled word of mouth, but as impressive as their armor is it's pretty clear they aren't actually all that, and the bosses clearly have the plan to deal with it. The mood along the defensive lines is significantly more cheerful now.

It would have been better if that had gotten all the local  ?birds? They’re not dragons or gryphons, not of any variety he’s ever heard of, but his experts were not particularly confident in any guesses as to what they were instead.

The first engagement was one sided, just like you always hope for in a clash with enemy elites, but they’d have to be spectacularly stupid to let him do it again. If he’s lucky, it’ll at least have been enough of a scare that they’re not in a hurry to try it again. If he’s not lucky, they’ll be angry enough to trust in whatever countermeasure the barbarian wizards can cook up, and his legion will get a chance to find out firsthand just how effective the barbarian air cavalry and wizards are. There’s little he can do to affect that decision, save through beseeching the gods.

But he can take steps to mitigate the results. His best wizards will rotate back from the front lines, to conserve their power and ensure they don’t fall victim to whatever has been targeting his dragons; lesser mages will have to suffice to handle any threats that emerge. And he can rearrange the lines, sending his subordinates to do the job wherever he lacks the information to decide properly. Some places took the abortive assault better than others, and depending on why will need to be either reinforced or abandoned. And he can make sure the wyverns know to expect water magic from their counterparts.

All told, it's a bit under five hours before the next major operation gets underway. Unlike the previous, it's not intended as a major push to regain territory or press to the portal, but that doesn't mean they've been slacking on the preparations. What can they observe about the disposition of the alien forces?

At night, while encamped? Well, that depends on how close they're getting.

Assume for the sake of argument that they're doing aerial recon, from mostly a couple thousand feet out but with a few closer flybys, and that said scouts have vastly better vision than any human but no ability to see through solid objects or otherwise get a good look at anything without a clear line of sight to the sky. As well as of course numerous groundbound observers at various distances from the front lines and a handful of interspersed security cameras still relevantly functioning. There are also teams of experts standing by, to hopefully get the most out of all the individual data points they can get their hands on, but that's the kind of thing that famously offers rapidly diminishing returns even when you aren't this confused.

(They don't have a satellite to supplement it, unfortunately, as reassuring as it would be as a way to supplement their potentially risky drones. There's an amount of civilizational capability that would allow them to relocate a specialized ground observation platform above an arbitrary city this quickly, but it's not one they had, and there hadn't happened to be any bizarrely convenient for the task).

Does that include sounds?

Yes, quite well actually - they've got actual audio recording on the ground, and can use laser mics to pick up more if it's being said somewhere convenient. But it's a bit hampered by the fact that they still don't have a working translation for the language, even if they think they've got some of the words from context clues.

In that case they can get a fairly good look at the location and disposition of most of the troops and sentries. The empire isn't trying to make it obvious, but they don't know what the attack surfaces they're dealing with look like, and even if they did it would be pretty expensive to fully prevent.

From the looks of things, they've got sentries along most of the ring, but they're not always densely packed - it's possible, maybe even probable, that dath ilan is missing some better hidden ones in some places, but that still leaves large areas inadequately watched if they can figure out which ones. The majority of the sentries they do have are scattered around the four big encampments near the current front lines - from the looks of things, their primary worry is a night attack on sleeping soldiers. Most of the people on guard don't look like baseline humans - the green-skinned soldiers, in large part - but there are some humans among them, and a scattering of other appearances. Comparing all the soldiers they can see to everyone they think has gone through the portal, minus battle casualties, suggests they probably are missing less than 10% of the enemy forces total, but it could easily be lower.

They'd like to get crosstabs on comparison to everyone they've seen before, to try and get an idea if any group isn't showing up where they can see, but there's a lot of detail missing for that. Most of the soldiers they can get eyes on aren't being observed in that much precision - enough to confirm there's a person there,  and in some cases which sub-variety they are, but it's easier to tell someone is there than to tell who it is - particularly so when technologies like thermal imaging enter the equation. They'll still sort and tag the appearances they can, though, in the hopes that an elite team of face recognizers can track any patterns when compared to earlier or future observations.

Elite teams of what?

Anyway, it's a reasonably well set up set of defenses. Not as good as if he could get his men back to the actually prepared buildings by the gate, but in addition to all the much-needed money he'd be giving up retreating that quickly would make him look weak, to his rivals learning about it later if not to his men, so needs must when Emroy drives.

It would indeed be quite inconvenient if they were trying to take out the enemy soldiers or slip an army past, especially nonlethally with only the capabilities they've already demonstrated. But in between the front lines and the inner fortress is a lot of empty space, practically undefended, and with a combination of night vision allowing them to move without lights, underground tunnels, and aereal overwatch they can slip through the lines and pass unimpeded to very nearly the inner citadel itself, hopefully without issue. They want eyes nearby the gate, in case any of their hastily assembled sensor arrays can actually get anything out of it that would give them a better understanding of how all the alien tricks actually work, and some special operations teams with them in case they end up needing it later. There's enough situations going forwards where additional vectors to scramble a rapid response with could be key, such as whatever took out the helicopters deciding that drones are fair game too. If all goes well, they won't get noticed.

...By which they mean, unless they're wrong about which possible blindspots happen to have observers and things also go in the bottom 10th percentile of expected outcomes, or else they're enormously wrong about something else major going on here. For some reason it felt important to specify they weren't trying to bet on a slightly weighted coinflip here, even though you'd really think that goes without saying when true lives are on the line and the conjunction of <all goes well> and <operation fails due to unexpected inputs currently trading at less than 2%> hovering around 96% would suffice to clarify if not.

He didn't actually do this, because he isn't completely insane. You could call him a bold general, inclined to risk defeat in pursuit of victory, and not be wrong, but it's not like being careful is a foreign concept. He's got the sort of caution that leads a general to wonder if this wide open valley they're walking into is an enemy ambush, and knows how to hedge against assumptions being wrong in the face of surprising enemy competence - a skill that has already served him well a few times this last day. But he doesn't feel the slightest temptation to leave his troops undefended against a threat that actually exists just for arbitrary paranoia, so his best secret observers are instead keeping watch to better raise the alarm around the places and people it would hurt the most to lose.

With that said though, the gate is his single most critical military asset right now, so in addition to the fortifications he's god several thousand troops stationed in the area and a handful of more esoteric defenses. 

In that case they'll abandon the approach of their first choice of forward base, and set up on the 14th floor of a high rise a two and a half blocks away from the gate fortress. It doesn't quite get them line of sight on the gate itself, but they can get over the walls, which is enough to get eyes on  sizable portions of the fortress interior. It's neither the best possible position for sniping nor for observation, but it's solid at both and far enough away they probably won't be noticed. 

Saderan military engineering is perhaps not optimized for cities where the buildings get this large, yes, and especially when it's not a one-off structure you can secure separately. In their defense making a fortress that can't be seen into by at least one of the surrounding skyscrapers would be a significant ask, and some quick examinations on demolition feasibility - 

Of course it's difficult to knock down a skyscraper, that would kill people! If your skyscraper was easy to knock down the insurance premiums to have it in a city* would bankrupt you inside a month.

*Doing it outside a city is much more feasible, especially since the land is also less expensive, but it's still not a common occurrence. Most people who build skyscrapers aren't in the business of getting them pointlessly knocked down, there being far easier ways to light money on fire; as a practice it is largely reserved for the sort of ultra wealthy supervillain-gendertrope who finds themselves unsatisfied with any lesser visual setpiece.

- made it clear that collapsing the larger buildings, in addition to being dangerous to anyone in the area, would be completely infeasible without a full team of siege wizards for each one.

Which, as little as they understand the alien capabilities, is not something they appear to have. It's good news for the team camping out in one now, and hopefully good news for the detachment that'll try and slip some autonomous monitoring equipment into closer structures.

It sure sounds like everything went as planned. Their next operation is in another 4 hours -

Actually, I think you'll find that it's our plan that comes up next. Eight hours after the last engagement, under the cover of darkness, the imperial dragon riders are returning to the sky for night operations.

It's a bit of a risk, and the typical rule is that you're not supposed to do that with your dragons, but there's good reason for it. For one thing, it should be significantly safer - it'll be harder to see the wyverns under the cover of darkness, and given how long it took between their arrival and the first attacks it seems unlikely that whatever countermeasures they had are something that can work tirelessly day and night at full efficiency. Secondly, this time they have targets - not a meandering search pattern over the city for important buildings or pockets of resistance, but actual troop concentrations, and with any luck whatever eyrie their counterparts flew off to after taking their own licking. They can leave cover, head in for the attack, and withdraw before there's time to organize a real response.

And, admittedly, it's not like the alternatives are amazing either. The riders resent the fact that they've had to hang around in hiding after taking a licking earlier, especially after enough time has passed for the initial fear to wear off. He can order them to refrain anyway, but it'll breed resentment and stress his authority over the borrowed units, which isn't worth it when he honestly really would like some air support against the barbarians. If they're willing to take the risks of pulling it off, best to ensure it's as helpful an operation as possible if they succeed.

The cover of darkness makes it a fair bit harder to spot the wyverns with the naked eye. Even night vision goggles don't solve that perfectly, though they're good enough that Civilization would have at least some advance warning. But what the darkness doesn't help with one iota is radar, and in fact they're aware of the flight almost as soon as they take off from their temporary aerie. 

The wyverns are objectively perhaps the most dangerous of the alien troops, at least counting only those they've managed to get a positive ID on, but by now that mostly means that dath ilan does not feel sufficiently secure to hold back and try to keep them alive. As soon as the draconic vanguard gets close enough, they're met with a coordinated barrage of wire-guided missiles against which their evasive maneuvres are useless.

How bad is it? They're prepared to press on in the case of casualties, but if it's devastating enough the rest of the vanguard might have second thoughts.

There is no rest of the vanguard. Not anymore.