My suggestions:
You might want to just copy the grill and a single headlight and put those on separate layers. You can then rotate the headlight and/or resize each part individually to your liking. On the asphalt supers you might want to rotate the lights to make them more level. On the dF model, with how Adam has the nose mapped... putting some angle into the lights generally makes them line up a bit more accurately. If I was to do these up for the SBS Phase 3 LM's, I'd likely rotate them back.
The other thing to consider is these are designed at 2048... if you are working on a 1024 x 1024 template you will have to resize them and will likely lose quite a bit of quality in doing so. When you resize any bitmap image, the more you do it (whether you scale any image down you will lose some, but you lose even more if you scale an image multiple times... i.e. 2-3-4 times vs. just once) the more you are prone to experience artifacting/chunking in the process of the scale. Also keep in mind that scaling generally works better in more round #'s, so going from 100% to 75%, 50%, 25%, is much better than going with 84%, 72%, 48%, and so on.
Bitmaps approximate vs. vector graphics which are more processor intense, but also generally scale to more finite #'s (which is why vector artwork is generally very clean and clear when scaled vs. bitmaps which tend to get fuzzy/hazy edges). To put it mildly, a bitmap graphic will tend to distort considerably the more you scale it... and since these are already scaled from high res imagery, you might in some cases be better off doing the process yourself if you aim to size these for 1024 x 1024. If you were to make a 2048 x 2048 asphalt Super LM template though, these could likely work much better (you are also better off resizing an entire 2048 x 2048 image to 1024 x 1024 for use in-game than you are resizing each individual layer to 1024 x 1024 IMHO). If you are wondering where some of the art I use is taken from... go look around on Leftlanenews.com and other automotive sites. The key with Leftlane, is when you open one of the image galleries... right-click on an image you want to save for editing, open it in a new window/tab in your browser, and then on the url... delete the ?= part after the .jpg of the url link and refresh the image. You'll most likely go from a smaller resolution image to something that's well over 1024px x xxx px. The larger the resolution base art you have to work with, the better. In some cases what I use to make the noses is from scratch... some I mix/match, and some is totally high res to start with. In those cases where it's scratch, you'll either have to try to salvage what I gave you or take a go at making it up yourself.
One process I'd recommend for cleanup around the edges of graphics is to bust out the pen tool (make a shape layer = my preference, I like the reusability of having a shape vs. a mere path) and trace around the outside of the various shapes. When done, go to the "paths" palette in Photoshop, at the bottom of the palette is a series of buttons. One is labeled "convert path to selection". Once you have it selected, there's an option in one of the menus called "inverse" which will flip the selection. Once flipped, I generally press the delete key on my keyboard on the active layer I want to clean and it will cut off any of the chunking that falls outside of the selection from the bitmap rescaling.
In case anyone is wondering... that's part of the process on how I make the noses to begin with. I make a vector shape layer for each part of the nose, whether it's a headlight decal or grill decal, and then when I copy out a piece of an original high resolution graphic, I will not only duplicate the original bitmap selection to my nose presentation .psd that I have a copy saved at for dF cars... but I also move the vector layer into the .psd and align it and link it with the original bitmap. That way when I scale or move the light graphic, the vector layer scales/moves with it. Once I have it finally placed, I then will go through the process of doing a selection once again, doing an inverse selection, and then deleting any of the fuzzies that lurk outside of the shape. That way if I ever decide to do a solid colored headlight and grill, I can just use layer effects in Photoshop and do a color overlay and the original graphic remains clean.
