The coming extension of cellular technology to base-stations in low-earth orbit (LEO) requires a fresh look at terrestrial 3GPP channel models. Relative to such models, sky-to-ground cellular channels will exhibit less diffraction, deeper shadowing, larger Doppler shifts, and possibly far stronger cross-cell interference: consequences of high elevation angles and extreme "sectorization" of LEO satellite transmissions into partially-overlapping spot beams. To permit forecasting of expected signal-to-noise ratio (SNR), interference-to-noise ratio (INR) and probability of outage, we characterize the powers of desired and interference signals as received by ground users from such a LEO satellite. In particular, building on the Shadowed Rician channel model, we observe that co-cell and cross-cell sky-to-ground signals travel along similar paths, whereas terrestrial co- and cross-cell signals travel along very different paths. We characterize SNR, signal-to-interference ratio (SIR), and INR using transmit beam profiles and linear relationships that we establish between certain Shadowed Rician random variables. These tools allow us to simplify certain density functions and moments, facilitating future analysis. Numerical results yield insight into the key question of whether emerging LEO systems should be viewed as interference- or noise-limited.