For the first time, the cloud radiative effect (CRE) has been characterized for the Arctic site Ny-Alesund including more than 2 years of data (June 2016 - September 2018). The cloud radiative effect, i.e. the difference between the all-sky and equivalent clear-sky net radiative fluxes, has been derived based on a combination of ground-based remote sensing observations of cloud properties and the application of broadband radiative transfer simulations. The simulated fluxes have been evaluated in terms of a radiative closure study. A good agreement with observed surface net shortwave (SW) and longwave (LW) fluxes has been found with small biases for clear-sky (SW: 3.8 Wm^-2; LW: -4.9 Wm^-2) and all-sky situations (SW: -5.4 Wm^-2; LW: -0.2 Wm^-2). When moving to monthly averages, uncertainties in the CRE are estimated to be small (~2 Wm^-2). At Ny-Alesund, the monthly net surface CRE is positive from September to April/May and negative in summer. The annual surface warming effect by clouds is 11.1 Wm^-2. The longwave surface CRE of liquid-containing cloud is mainly driven by liquid water path (LWP) with an asymptote value of 75 Wm^-2 for large LWP values. The shortwave surface CRE can largely be explained by LWP, solar zenith angle and surface albedo. Liquid-containing clouds (LWP>5 gm^-2) clearly contribute most to the shortwave surface CRE (70-98%) and from late spring to autumn also to the longwave surface CRE (up to 95%). Only in winter, ice clouds (IWP>0 gm^-2, LWP<5 gm^-2) are equally important or may even dominate the signal in the longwave surface CRE.