Nitrogen (N) is the primary limiting nutrient for forest production. Therefore, understanding how environmental factors affect N transformation rates is essential for the provision of sustainable ecosystem services. Because these factors are interlinked, it is important to consider direct and indirect structural relationships to better understand the factors contributing to N transformations. In this study, we analyzed the structural cause-effect relationships surrounding N transformations by structural equation modeling using a database containing net and gross N transformation rates and related soil chemical properties from 38 sites across the Japanese archipelago. The average net N mineralization and nitrification rates in the Japanese forest soils were 0.62 +/- 0.68 and 0.59 +/- 0.65 mg N kg(-1) d(-1), respectively, and gross N mineralization and nitrification rates were 4.22 +/- 3.59 and 0.98 +/- 0.68 mg N kg(-1) d(-1), respectively. Compared with previous large scale studies, net and gross N transformation rates in Japanese forest soils were considerably diverse despite their relatively small land area and were representative of temperate forest ecosystems. Structural equation modeling analysis showed that net N transformations were directly affected by gross N transformations, which in turn were significantly and directly affected by soil organic matter contents. Soil organic matter was significantly affected by organic layer amount, tree species and soil group. The effect of soil group was the greatest among these factors, suggesting that soil organic matter contents in Japanese forest soils were mainly influenced by soil parent materials. This was especially evident for Andosols, which are derived from volcanic sediments and contain large amounts of soil organic matter leading to high N transformation rates in the Japanese forest soils. Among the factors related to organic layers and mineral soil layers, soil organic matter content and organic layer amount, which represent substrate availability, had significant effects on gross and net N transformation rates. However, by refining the scale of the dataset using soil groups/soil parent materials, the influence of substrate quality and soil chemical properties on N transformations was suggested. From the current dataset, it was indicated that soil parent materials were the most important factor controlling the pattern of N transformations in the soil of Japanese forest ecosystems. This conclusion should be repeatedly refined considering the spatial distribution of factors such as climatic conditions and forest types with additional site datasets obtained from future surveys. (C) 2015 Elsevier B.V. All rights reserved.