2 -Ka-band link performance is dominated by weather effects, both in terms of atmospheric attenuation and atmospheric noise temperature increase. Weather effects are inherently statistical, and Kaband weather effects have higher uncertainty compared to the S-band and X-band counterparts. As such, the Gaussian assumption that simplifies statistical link analysis for S-and X-bands might not be valid, and computational-intensive convolution of probability density functions are needed for link analysis. The weather effects, whether individual or aggregated, can be easily measured but difficult to characterize as parametric probability distribution functions. To simplify the analysis, the current statistical link analysis approach uses the worst-case weather loss as a deterministic value to evaluate the link performance, which results in a conservative link design. Using sound statistical principles, this paper describe a simple link analysis approach that integrates the Gaussian components of a link directly with the empirical Ka-band weather data from direct measurements This approach removes the artificial conservatism of the current approach, and yields a lower and tighter upper bound of the received power required to close the link in a statistical sense. We illustrate this technique and compare with the current approach on link analysis using the antennas and weather measurements of the three DSN sites in Goldstone, Canberra, and Madrid.
NomenclatureP T N 0 = total power signal-to-noise ratio EIRP = effective isotropic radiator power G/T = gain over system noise temperature k = Boltzmann's constant (1.38 x 10 -23 J/K) L s = space-loss T sys = system noise temperature T atm = atmospheric noise temperature L atm = atmospheric attenuation θ = elevation angle CD = weather availability (in %) T p = mean atmospheric temperature L tot dB = total weather loss (in dB) m = mean σ = standard deviation l CD = a deterministic number that corresponds to the value of L dB tot (θ ) such that Prob[ L tot dB (θ ) ≤ l CD ] = CD