Phenomenology (from a non-philosophical point of view) is commonly applied to the field of particle physics, where it forms a bridge between the mathematical models of theoretical physics and the results of high-energy experiments. It is also used in some cases in which there are no existing theories to describe the observed experimental data, such as in the physics of condensed matter and plasma physics. As a result, I think that the main difference between microscopic level and phenomenology is given by the frontier in which the Planck's constant begins to have a significant influence to describe the phenomenon under study. If the elementary unit of a given phenomena is not under the influence of Planck's constant (i.e.: it shows neither wave-particle duality nor uncertainty in the Heisenberg meaning of the word), then the phenomenon is of microscopic nature. On the contrary, if it displays the above-mentioned properties (quantum nature, wave-particle duality, and Heisenberg uncertainty), the phenomenon has overcome the microscopic level until reaching the level of quantum physics phenomenology.

Even if I have appreciated the Racardo answer, I think that there is something more fundamental than the quantum effects in determining the difference between the "individual" and "many body" regimes, independently on the size of "particles" (here used as a synonymous of individuals). Many systems in nature presents this kind of behaviors. Let us think of groups of humans and the different behaviors of the mass with respect to the rationality of the individuals.

If the density becomes large enough, the individual wills disappear and self- organization or other global behavior appears. It seems a phase transition.

There is an Italian physicist, Giorgio Parisi, investigating this phenomena in a general way, applying his results to different systems.

I can give a simple example to outlay the differences between microscopic treatment and phenomenological modelling.

Example of an inflated balloon with gas particles inside. When you apply pressure radially inward, the gas molecules will gain energy and there will be some heat dissipation as well. You can use microscopic atomic level treatment to model this, though you may also need statistical thermodynamic techniques to predict a future of the gas ensemble that will have some probability attached to it. The probabilistic nature is an inherent part of the dynamics of the process, almost accepted by the scientific community. This is a microscopic treatment. Note: Statistical thermodynamics is in general a technique to predict the future of a complex system. It cannot be called phenomenological unless one does a lot of hand-waving in the treatment. Statistical thermodynamics has physics at its core!

However one can also predict the gas balloon behavior using a combination of spring and a Newtonian dash-pot, the spring will represent the lossless process and dash pot will represent the heat dissipation. However, there is neither a spring inside the balloon nor any viscous dash-pot, yet such a modelling can be used the study the "approximate" behavior of the system. This is a phenomenological approach, though very useful in studying complex system has often been found challenging to connect it with the real physical parameters of the process.

Many thanks to all of you to have spared some time and commented on the subject of microscopic vs phenomenology (macroscopic?). Gianpiero Colonna's comment "It seems a phase transition" is what matches my mind on the matter. I think, may be it needs correction, that it takes substantially many microscopic entities to act at once in a certain event to create an effect that we as human beings are able to sense. Therefore, I believe phenomenon is an event large-scale enough for our faculties to realize it while microscopic activity is beyond where it may let us get hold of it. When is it that the scale of the visible transitions to the scale of the invisible and is it linear, nonlinear or logarithmic are the tantalizing nuts yet to be cracked!