Hi Tarek, your question is very interesting. Can you please apport more precision on the greenhouse type, cover (plastic or other), conditions... these elements can influence the response.  

The greenhouses are plastic tunnels characterized by circular sidewalls (50 meters length and 8-9 meters wide). The arcs constituting the greenhouse structure are made from metal (galvanized iron). To ensure ventilation, two doors are created on both sides of the tunnel (front and back).

Dear Tarik,

Here are some suggestions about how to address your question:

- The first issue to settle is "How accurate should your ETP calculation be?" In other words, what is the largest error you can tolerate in your estimation that will not invalidate the entire exercise? If you do not have such a quality criterion, then any model and any answer will work...

- The next issue is "How will you estimate the actual accuracy of your estimations?" Do you have an independent way to measure ETP, that does not involve this equation? Since you are working with a greenhouse, I assume your plants are held in pots. So it should be reasonably easy to place some of those pots on accurate scales to measure, continuously or at regular intervals, the mass of the pots + soil + plants. Each time you water them, the mass reading should increase by exactly the amount of water applied, and then the total mass should evolve in time to reflect both the growth of the plant itself (which will absorb CO2 from the atmosphere) and the loss of water through ETP. Of course, this also may involve estimating the mass of the growing plant, but this should be feasible with separate measurements on multiple plants of the same species, grown under identical conditions: you will need to establish allometric relations between plant size and mass, for instance.

- Assuming you have a clear accuracy goal and an independent empirical method to assess the actual ETP, you should be able to establish whether the estimations generated with your formula bear any resemblance to your measurements, for those potted plants that are growing on scales.

- You will then be in a position to decide whether that formula is good enough for your purpose. In fact, if you have a sufficient number of observations, and are confident in their own accuracy, you might even be able to "tune" the equation for your purpose and guarantee a better fit between the model and the observations.

- Lastly, if you find that this approach is too unreliable, you will still have the option to adopt and evaluate, using the same approach, another, perhaps more complex or more accurate model.

Good luck in your investigations. Michel.

Very interesting and typical question as well situation, few points coming to mind

1. Radiation and aerodynamic terms, contributing to potential/reference ET,  

2. radiation balance modified under greenhouse, need to include radiation comp of canopy environment

3. aerodynamic component, since partial ventilation exist need to be considered, but how needs to be debated

4 Priestley and Taylor's method which does approximation for aerodynamic component through radiation balance calculation only could be used

5. Not difficult in your situation, as your K value which you are taking as 0.65 needs to be calibrated, but the coefficient may change from greenhouse to greenhouse 

6. Calibration and validation is a bit tedious under this situation

regards