I think the answer added is nice. You already have the distribution of power over the wavelength. But with the suggested method you can calculate p/deltalamda at a given wavelength.

Thanks Michael and Prathik, this solution is seems to me a perfect one.