Аннотация:The demography of Picea abies trees was studied over a period of about 30 yr on permanent plots in six forest types of an unmanaged forest located in a forest reserve of the Southern Taiga, NW of Moscow. This study encompassed a broad range of conditions that are typical for old growth spruce forests in the boreal region, including sites with a high water table and well drained sites, podzolic soils, acidic soils and organic soils. At all sites stand density, tree height, breast height diameter and age has been periodically recorded since 1968. Tree density ranged between 178 and 1035 trees ha−1 for spruce and between 232 and 1168 trees ha−1 for the whole stand, including mainly Betula and Populus. Biomass ranged between 5.4 and 170 tdw ha−1 for spruce and between 33 to 198 tdw ha−1 for the whole stand. Averaged over a long period of time, biomass did not change with stand density according to the self-thinning rule. In fact, on most sites biomass remained almost constant in the long term, while stand density decreased. The study demonstrates that the loss of living trees was not regulated by competitive interactions between trees, but by disturbances caused by climatic events. Dry years caused losses of minor and younger trees without affecting biomass. In contrast, periodic storms resulted in a loss of biomass without affecting density, except for extreme events, where the whole stand may fall. Dry years followed by wet years enhance the effect on stand density. Since mainly younger trees were lost, the apparent average age of the stand increased more than real time (20% for Picea). Average mortality was 2.8 ± 0.5% yr−1 for spruce. Thus, the forest is turned over once every 160–180 yr by disturbances. The demography of dead trees shows that the rate of decay depends on the way the tree died. Storm causes uprooting and stem breakage, where living trees fall to the forest floor and decay with a mean residence time (t1/2) of about16 yr (decomposition rate constant kd= 0.042 yr−1). This contrasts with trees that die by drought or insect damage, and which remain as standing dead trees with a mean residence time of 3–13 yr until they are brought to ground, mainly by wind. These standing dead trees require an additional mean residence time of about 22 yr for decay on the ground (kd= 0.031). In conclusion, we demonstrate that, rather than competitive interactions, it is climate extremes, namely drought, rapid changes of dry years followed by wet years, and storm that determine stand structure, biomass and density, which then affect the net exchange with the atmosphere. The climatic effects are difficult to predict, because the sensitivity of a stand to climate extremes depends on the past history. This may range from no effect, if the stand was recovering from an earlier drought and exhibited a relatively low density, to a total collapse of canopies, if drought reduces stand density to an extent that other climatic extremes (especially wind) may cause further damage.