Put all the attempts at different things that you tried here.
Doesn't appear to be Vignere. Performed key size analysis described at these instructions (last paragraph.),[3278,3299,3298,2838],[3288,3294,3296,2472],[4516,1206,708,1820]] this Wikipedia article], breaking the ciphertext into k columns where k was the suspected key length, and computing the average of the index of coincidence (IoC) of each column. Blocks of ciphertext from size 50 to 5000 were also considered separately, in case the runes are enciphered using different keys at different points in the text. Observed max IoC < 1.15 (compare to 1.6 for some sample plaintext, which theoretically should be the same if k were correct and the text were enciphered using Vignere.)
Python script here.
In general, it's not likely any cyclicly-repeating polyalphabetic cipher, since those are special cases of Vignere.
No hits on searching for an autokey cipher with key length 3 to 20, using a modified version of the python script mentioned in this page
Could use Hill cipher:
The other base-60 text was cuneiform, and appeared as a 2x2 matrix. Tried that as the key and deciphered using Hill. No results, and the cuneiform was later revealed to be "3301" and "1033" in base 60.
Tried to use the 4x4 matrix as the Hill cipher key, without success.
In both of the above, frequency analysis of n-length binary substrings was done to check for statistically significant results from decryption, using this script.
Attempted Error-Level Analysis of the jpgs. No results so far, up to image 4.