That seems like reasonable starting point, there will always be uncertainty related to the type of tree, how it is maintained, and how old it is for the simulation year. Also the size one chooses to model it, it isn't always clear what the shading surface should look like, in some cases I modeled three intersecting octagons on x,y, and z. They were large trees close to the building so horizontal or vertical surface by itself didn't seem right.

There's an updated version in 1995 (published 1997) from McPherson, they use 0.7 summer and 0.175 winter.

Hi Julien, can you please give us the complete reference of the updated version that you mentioned? there are any new one maybe?

Another question, isn't should have higher values of transmittance the deciduous trees in winter, and not the opposite? i understand that from https://bigladdersoftware.com/epx/doc..., where ... for deciduous trees that have a higher transmittance in winter than in summer...

The study talks about a tree shading factor, which would be the opposite of the transmittance. The study I linked to has a few references to McPherson, the one that matters is called "Simulating tree shadow patterns for building energy analysis", and you will find a 1998 study called "Simulation of tree shade impacts on residential energy use for space conditioning in Sacramento" (cf here) that mentions the 0.7 summer and 0.175 winter factors I put above. McPherson appears to have published a bunch of studies after that too