An empirical formula is developed for predicting monthly sea level strontium 90 fallout (F) in the northern hemisphere as a function of time (t), precipitation rate (P), latitude (ϕ), longitude (λ), and the sea level concentration of stronium 90 in air (C): F(λ, ϕ, t) = C(t, ϕ)<v d(ϕ) + vw(λ, ϕ, t)>, where vw(λ, ϕ, t) = a(ϕ)<P(λ, ϕ, t)/Po>b(ϕ) is the wet removal, vd(ϕ) is the dry removal and P0 is 1 cm/month. The constants vd, a, and b are determined as functions of latitude by fitting land based observations. The concentration of 90Sr in air is calculated as a function of the deseasonalized concentration at a reference latitude (C¿ref), the ratio of the observations at the latitude of interest to the reference latitude (R), and a function representing the seasonal trend in the air concentration (1 + g): C¿(t, ϕ) = Cref(t)R(ϕ)<1 + g(m, ϕ)>; m is the month. Zonal trends in C are shown to be relatively small. This formula can be used in conjuction with precipitation observations and/or estimates to predict fallout in the northern hemisphere for any month in the years 1954 to 1974. Error estimates are given; they do not include uncertainty due to errors in precipitation data. The cumulative North Atlantic 90Sr input predicted by the model for 1.25¿N to 66.25¿N decay corrected to December 1972, is 1920 +160-237 kCi, compared to an inventory estimate of 2440 +70-265 kCi from oceanic observations. It is suggested that the discrepancy is due to a more efficient scavenging of small particles over the oceans than over lan