Constructs for describing predictions of water quality?

In the Digital-Water.city project, a use case is using statistical models for predicting the water quality in the river. The predicted measurements should be associated with different percentiles with cardinality 1..* (e.g., a predicted value of 0.3 (in the range 0-1) at percentile 50, another prediction of 0.4 at percentile 97.5) and it should be possible to describe a conclusion from the prediction stating the overall water quality using an enumerated value ("Excellent", "Good", "Sufficient", "Poor"). A sketch of a possible solution is enclosed: