The coupling strength (CS) between land and atmosphere in global scale, especially, influence of the land surface states (soil moisture, soil temperature, canopy temperature, canopy water, etc) to precipitation variability has been quantified with atmospheric general circulation models (AGCMs). The authors participate the GLACE (Global Land-Atmosphere Coupling Experiment), which is called for by Koster et al. In GLACE, a dozen climate-modeling groups have performed the same highly controlled numerical experiment to quantify CS for precipitation in boreal summer. In this study, CS for several atmospheric variables (surface temperature, evaporation, air temperature, specific humidity, relative humidity and precipitation) is quantified with the CCSR/NIES (Center for Climate System Research and National Institute for Environmental Studies) AGCM; especially CS for the precipitation is focused over the land in global scale. The result suggests that the degree of CS for each variable has much difference. Large CS is found for variables, which exist near the land surface. Especially, CS for the precipitation is small compared with others. This can be mentioned the difficulty of weather forecasting for the daily precipitation variability (6 day totals). However CS for the precipitation is shown relatively large over the North America, Central Eurasia and Central Africa. In these regions, the land surface anomalies can strongly contribute to the daily precipitation variability compared with other regions. The reasons why CS for the precipitation is largely evaluated over these regions are also focused in this study. Constraint factors for determining the degree of CS can be existed in the land or the atmosphere. The results suggest that CS for the precipitation is very sensitive to (a) the dryness conditions over near the land surface (b) and the thickness or the degree of activity of the atmospheric boundary layer. (a) Firstly, the relation of CS for the evaporation and the dryness condition (relative humidity) near the land surface is compared over the land in grid scale. As a result, CS for the evaporation has the maximum value to the relative humidity (0.2-0.4) near the surface. Namely, CS for the evaporation is large over the arid region or the temperate zones. On the other hand, CS for the evaporation is small over the extremely humid or the dry conditions over the near land surface. Over the extremely humid condition, much soil moisture can be stored in the soil layer, so the capacity to store the evaporated soil moisture can be large in the atmosphere. However, the atmosphere near the land surface has also much water content in the extremely humid condition. Therefore, the soil moisture cannot be evaporated much into the atmosphere. In the other reason, the soil moisture, which evaporated from the soil layer, cannot be dominant for the total amount of the water content in the atmosphere. In the extremely dry conditions near the land surface, the atmosphere has much capacity to store the water content. However, the content of the soil moisture in the soil layer is small for the reason of the extremely dry conditions. As a result, CS for evaporation can be shown small value. (b) In vertical scale, CS for the air temperature or the specific humidity is large within the atmospheric boundary layer. In other words, CS for these variables largely decreases within the atmospheric boundary layer. As results of (a) and (b), CS for precipitation can be strongly influenced by two processes, the dryness conditions near the land surface and the state of the atmospheric boundary layer. Dividing the precipitation type into the large-scale condensation type (large-scale precipitation) and the convective cloud type (convective precipitation), large CS for the convective precipitation is more largely distributed than that for the large-scale precipitation in global scale. The convective precipitation is generated in horizontally small scale compared with the large-scale precipitation. This is the reason why the land surface anomalies can strongly contribute the convective precipitation. The authors analyzed the influence of the land surface anomalies to the precipitation variability for several time scales (1 day, 1 week, 2 weeks, 1 month and 3 months) by using not CS but other statistical approach. The result suggests that the influence of the land surface anomalies to the precipitation in 3 months case is large over many regions in global scale.