That's a good question.  The trends by themselves do not give you probabilities, but they do shift the odds toward one category (above or below normal) and away from the other.  You can imagine that we have a probability distribution of seasonal temperature or precipitation at some location.  The average of the distribution determines whether the probabilities are tilted toward above or below average.  The width of the distribution, which is determined by upredictable weather noise and can be obtained from forecast models or historical data, determines what the probabilities are in each category.  If the trend is positive, then the average of the distribution is shifted to the right relative to the base period, and the probabilities for above normal increase while the probabilities for below normal decrease, but the width of the distribution does not change.  Therefore, the trends in combination of the shape of the distribution, which is determined by other factors, result in changes in probabilities.