I've been looking at HiRISE images, and this description says that this crater is new, which is why we can see the bright ejecta. How long does it typically take for the bright ejecta to erode away? What does "new" mean in this context? Was the crater formed a few days before the picture? A few years? Centuries?
To directly answer your questions before providing context:
- It takes "awhile" for brightness differences of ejecta to disappear on Mars, but unlike the Moon where it's around ~1 billion years, on Mars it's more like decades to centuries due to the active surface/atmosphere processes.
- The crater in that image formed some time between 04 June, 2010, and 12 November, 2020.
- In that particular image, the crater was formed at least 4.5 months before the image to which you linked was taken.
More Descriptive Answer
Change-detection on Mars goes back a long time, and once we started to get high-resolution cameras in orbit that returned data for a fairly long time (years), we started doing change-detection with respect to searching for new impact craters.
One of those cameras is the Context Camera (CTX), which is on the Mars Reconnaissance Orbiter, which has been returning data since 2006. CTX returns images at roughly 5–6 meters per pixel, is about 5000 pixels wide (≈30 km wide), and up to about 50,000 pixels long (≈300 km). It's returned nearly global coverage since 2006, and a lot of repeat coverage. While it's rare that CTX can resolve craters that are newly formed because they are often very small, it can resolve the bright or dark ejecta that is a few times larger than the crater itself.
Ingrid Daubar, for her doctoral dissertation around 2013, did an extremely detailed analysis of overlapping CTX images at the time, and she identified around 200 new impact crater that had formed based on changes between one CTX image and another. Many of these were targeted later by the HiRISE camera (≈25 cm per pixel) for more detailed analysis.
Ingrid remains the primary person who has done this work for Mars, and just last month presented new work with her latest database of 1,203 new craters: Those that formed some time between 2006 and 2022.
The crater you linked to in particular is on CTX image N12_067017_1709, taken 12 November, 2020, while it is not on CTX image B22_018068_1691, which was taken 04 June, 2010. I couldn't find any other CTX images that cover that location taken between those two times, so the crater must have formed in that ≈10.5-year period. After it was found, there are three more CTX images taken of just that area, starting 2 Earth months later and going through 4 Earth months later. Two of those were context images for the HiRISE follow-up that took a stereopair, starting 30 January, 2021, and then 12 March, 2021 (the latter being the one you linked to); the other CTX image is from 02 March, 2021.
So, that's why I can say that the crater in the image to which you linked formed at least 4.5 months before that image, since you linked to the second HiRISE image, but it could have formed up to ≈10.8 Earth Years before that image (any time after that June 2010 image was taken).