Forward error correction for an atmospheric noise channel

Two Markov chains are employed to model the memory of the atmospheric noise channel. It derives the transition probabilities for these chains from atmospheric noise error processes that were recorded at 306 kHz. The models are then utilized to estimate the probability of codeword error, and compares these estimates to codeword error rates that are obtained directly from the recorded error processes. These comparisons are made for the Golay code with various bit interleaving depths, and for a Reed-Solomon code with a variety of symbol interleaving depths.