Edited a bit (8:55PM):
Hi, disclaimer I'm a current undergraduate who also transferred from CC to UCSB, less successful (but not hopeless and ideas of success vary widely anyway) and applying to grad school next year.
I have many thoughts on this subject, but I'd reiterate the above that it probably won't matter too much. To answer your questions directly:
1. How they will look at it will vary from school to school. Obviously it's far better to have more research experience than less, most of all, and going to a cc will likely severely limit that. At least during the school year and summer research is difficult to obtain, but luckily there are many summer research programs that reserve spaces for cc/liberal arts college students (like those REU's so in a way it could be "easier" because you're competing with other cc/liberal arts college students, as opposed to university students).
2. Personally it felt like a train hit me, in terms of change in difficulty in classes, plus the semester to quarter system change was intense as well. Specifically for UCSB it is a bit weird in that its probably the easiest top ten in physics to get into for undergrad, but they basically "super-train" an already stellar group of students in a "research honors physics major" (CCS Physics, google will explain it better) that you have to compete with in upper-division classes, on top of the other regular physics majors that have had 2 years of lowers division physics classes tailored to preparing them for physics upper division classes at UCSB as well. So you're "competing" against students who have arguably 1 to 2 higher levels of intensive preparation than you do. However my performance was largely my fault (of course) since I did not prepare as well as I could have in lower division and have been trying to catch up ever since. But on the flip=side I am also doing a couple of grad courses in my 2nd year here, >1 years worth of computational/theoretical research under an undergrad fellowship with a reputable professor/grad student "theory group" (including over the summer), and will likely submit something for publication soon, so that's just my experience.
3. I answered more or less your research question in 1; in short it'd probably help to get an REU, but hurt in ease of getting research opportunities in general. In my CC (and possibly nearby state school) there was little to no Physics research and it was very difficult to obtain research in general in my hometown, so I started research my winter quarter at UCSB. I suppose my main question is which UC you got into to help answer this more in-depth.
Personally if you got into any of UCI, UCLA, UCSD, UCD, UCSB, UCSC maybe UC Riverside, I'd just go. Although what you actually want to major in and if you actually still want to go to graduate school eventually has a high chance of changing, you would probably be better off going to just any tier 1 research institution that's doing physics research you can participate in and enjoy, especially if you stick on this route. Especially if you get into UCSB CCS Physics, believe me I was a die hard "transfer to UC Berkeley or bust" kind of guy and there was nothing you could have told me if I actually got into UC Berkeley that would have stopped me from going. But really, go to a university you have a higher chance of getting all the stuff grad schools care more about like research experience, great letters, higher grades (especially for grades closest to application date, I imagine my CC grades weigh so much less than my UC grades and while your GPA "resets" when you transfer, that can be a good or bad thing), maybe publications, and etc.. The weight of all of these vary from graduate admissions committees of course, but are generally the most important from what I gather, you can see for yourself on the application results thread that it matters farrrr less what undergraduate institution you actually went to than what you accomplished there.
This unfortunately may mean going to say UCSB or UCD over UC Berkeley or UCLA if you feel your GPA will get wrecked (which it has a higher chance to, and no one feels their GPA would get wrecked initially so this is probably a moot point. Anyway, the UC's are probably all pretty similar in GPA wrecking ability) or if it would be harder to get a satisfactory research position as soon as entering the university. Yes it costs more to go to UC right away than to transfer (but to be honest you don't know if you'd get into the same or better places when you get to the transfer date and lots of people get "stuck in cc"), but I am of the (possibly regretful) mindset that this is an investment and if you have a higher chance of achieving you goals by taking these opportunities, despite the debt (which you can pay off if you succeed), I'd go to the UC over transferring.
So, if you get into say UCSB CCS Physics or think you can transfer into CCS Physics that is what I would do hands down looking back, and I'd do it over UC Berkeley and hell even over Stanford/MIT/Harvard probably (the program is that good). Otherwise, just do a bit more research into say STEM Summer/Quarter research opportunities/programs that are offered at the university, I'd imagine all of the UC's are good for these, and perhaps at the "lesser" UC's there's less competition for opportunities so you can focus on improving yourself and your application. I know for UCSB it's probably a bit easier to get into research earlier in your academic career thanks to an assortment of STEM research programs (although a lot are geared towards underrepresented students if that's a problem for you).
Hope that helps.