Dear authors,

I very much appreciate your efforts to carefully take into account the constructive comments of the referees, most of which were very accurate. This allowed the paper to be much more explicit about the drastic changes induced by non-Gaussian Levy stable forcing in an Energy Balance Model.

However, I noted that:
- While referee #2 only pointed out that physics and mathematics communities use respectively the terms “universal multifractals” and “Lévy multiplicative chaos”, you presented the latter “as a more mathematically tractable alternative” than the latter. Fortunately, there is no such claim in both cited references (Fan, 1997; Rhodes et al., 2014). Indeed, the expression “more mathematically tractable alternative” is used once by Rhodes et al. (2014) for the (discrete in scales) cascade model of Mandelbrot (1974) compared to the (continuous in scales) lognormal cascade model of Mandelbrot (1972). Both models do not deal with Levy generators at all. I believe there is no difficulty in eliminating this erroneous opposition, being understood there is no single way to define the Levy generator at finite resolutions, just as the Dirac measure can be defined by various limiting functional processes.

- A more complex issue is related to the question raised by the referee #2 about asymmetric or. symmetric Lévy noises. While I understand your arguments that this will introduce new complexity, more precisely add a new dimension to the parameter space and make it more difficult to explore, the referee’s arguments to drastically reduce the negative values of irradiance with the help of extremely asymmetric Lévy noises cannot be easily ignored. I therefore suggest to mention them in the conclusions as such for future work.

Since the above suggestions correspond to limited corrections, I am pleased to state that your paper is to be published subject to technical corrections.

Best regards, Daniel