Markov chains

What is users' next best alternative to the use of dynamical seasonal predictions?