Model for interdependant multivariate time series

Hello all,

What would be the apropiate way (or model) to forecast on a multivariate time series where the variables are ordered and each one have dependancy from the previous ones, but not the following ones.
For example the level of a river, lets say there are three ports on the same river, I have one measure per day per port, and I want to predict the measure for tomorrow for the last port.

Thank you very much!

© Copyright 2013-2019 Analytics Vidhya