Abstract. Our objectives were to identify factors related to temporal variation of soil CO 2 efflux in a boreal pine forest and to evaluate simple predictive models of temporal variation of soil CO 2 efflux. Soil CO 2 efflux was measured with a portable chamber in a Finnish Scots pine forest for three years, with a fourth year for model evaluation. Plot averages for soil CO 2 efflux ranged from 0.04 to 0.90 g CO 2 m −2 h −1 during the snow-free period, i.e. May-October, and from 0.04 to 0.13 g CO 2 m −2 h −1 in winter. Soil temperature was a good predictor of soil CO 2 efflux. A quadratic model of lntransformed efflux explained 76-82 % of the variation over the snow-free period.The results revealed an effect of season: at a given temperature of the organic layer, soil CO 2 efflux was higher later in the snow-free period (in August and September) than in spring and early summer (in May and June). Regression coefficients for temperature (approximations of a Q 10 value) of month-specific models decreased with increasing average soil temperatures. Efflux in July, the month of peak photosynthesis, showed no clear response to temperature or moisture. Inclusion of a seasonality index, degree days, improved the accuracy of temperature response models to predict efflux for the fourth year of measurements, which was not used in building of regression models. During peak efflux from mid-July to late-August, efflux was underestimated with the models that included degree days as well as with the models that did not. The strong influence of the flux of photosynthates belowground and the importance of root respiration could explain the relative temperature insensitivity observed in July and together with seasonality of growth of root and root-associated mycorrhizal fungi could explain partial failure of models to predict magnitude of efflux in the peak season from mid-July to August.Correspondence to: S. M. Niinistö (sini.niinisto@iki.fi)The effect of moisture early in the season was confounded by simultaneous advancement of the growing season and increase in temperature. In a dry year, however, the effect of drought was evident as soil CO 2 efflux was some 30 % smaller in September than in the previous wet year. Soil temperature was a good overall predictor of soil CO 2 efflux, possibly partly because its apparent effect was strengthened by many environmental factors and ecosystem processes that varied in concert with its variation. However, the consistent underestimation by the predictive models for the peak season corroborates recent findings concerning the importance of seasonal changes in carbon inputs to processes producing CO 2 in soil.