If you create a response surface or a full factorial designed experiment, you will test different combinations of factors and levels. Suppose you have 2 factors, time and enzyme conc at the beginning of the experiment. Suppose you have 5 levels (settings) for each factor, T1, T2, T3, T4, T5 and C1, C2, C3, C4, C5. You can test every combination of factor and level or test certain subsets of those combinations. That will give you the experiment you are looking at. 

It doesn't make sense that the extent of enzymatic degradation should decrease after reaching a maximum point. Even if the reaction is reversible, it should just reach a maximal amount of degradation that is less than complete. The only way I can think of to explain this graph is that the researchers were monitoring the appearance of a specific degradation product, rather than disappearance of the original substance, and this degradation product once formed was itself degraded into a second product.

Thanks Adam. I checked the paper again. They exactly discussed degradation efficiency of original substance.