『Abstract
　Lithology is an important characteristic of the terrestrial surface, whose properties influence chemical weathering rates. Specifically non-silicate minerals may contribute significantly to the weathering derived fluxes from silicate-dominated lithological classes. The Japanese Archipelago consists of predominantly silicate-dominated lithologies with a high proportion of volcanics. However, the spatially explicit representation of chemical weathering rates remains difficult for such a large region, because many controlling factors on chemical weathering rates are correlated with each other. Due to the spatial heterogeneity of lithology, a multi-lithological model approach to estimate spatially explicit chemical weathering rates for unmonitored areas is applied here. To achieve this, hydrochemical data of 381 catchments are used to train a set of models, recognizing the contribution of a variety of proposed factors influencing chemical silicate rock weathering rates (CSRWR: cations plus dissolved silica flux). The monitored catchments cover 〜44％ of the Japanese Archipelago. Cation chemical weathering rates (excluding Si) are linearly correlated with CSRWR and show outliers if basic volcanics or pyroclastic flows are present due to increased silica release rates.
　Lithology and runoff are identified as the strongest predictors for chemical weathering rates. Temperature and gradient of slope are of less relevance for the regional scale prediction while further proposed factors like soil properties or land cover are not identified as major predictors. Latter findings are partly attributed to geodata quality, low availability of parameter values as well as spatial correlations of proposed controlling factors with lithology or runoff.
　The calculated average CSRWR of the Archipelago is 〜25 t km^-2 a^-1 and ranges from 5.9 to 107 t km^-2 a^-1 in monitored catchments. Weathering rates per lithological class as a function of runoff can be grouped into three classes: a) pyroclastic flows showing the highest chemical weathering rates; b) alluvial deposits, mixed sediments and basic to intermediate volcanics with medium rates; and c) metamorphics, siliciclastic sediments, acid volcanics, and plutonics and unconsolidated sediments (other than alluvial deposits), showing the lowest rates. The recognition of lithogenic sulfur would add 9.7％ to CSRWR of considered catchments. Results suggest that the lithological classes acid volcanics and unconsolidated sediments contribute above average to the sulfur fluxes. Possible biases of this observation are discussed.
　The contribution of Ca-fluxes from non-silicate calcic minerals (named Ca-excess, Ca-fluxes in addition to silicate Ca-fluxes) is about 10％ of the CSRWR on average and is attributed by a wide value range. The calculated ratio “Ca-excess to total Ca-fluxes” from chemical weathering averages around 62％, 75％, 56％, 83％ and 84％ for the lithological classes acid plutonics, metamorphics, siliciclastic sediments, mixes sediments and acid volcanics, respectively. This suggests a major Ca-contribution from non-silicate calcic minerals for these lithological classes. Phosphorus release from rocks due to chemical weathering is estimated to between 1 kg P km^-2 a^-1 and 390 kg P km^-2 a^-1. The P-release patterns in dependence of runoff per lithological class are different from CSRWRs due to differences of applied P-content in rocks. The identified spatial P-release pattern suggest that the consideration of dynamic and spatially resolved P-release rates by chemical weathering might improve ecosystem studies. Later findings may be of importance for analysing the influence of P-release from rocks on the climate system via ecosystem functioning on geological time scales. A first application of the P-release model to the global scale suggests an annual release of 1.6 Mt P (13.8 kg P km^-2 a^-1) by chemical weathering of silicate dominated lithological classes (excluding carbonate sedimentary rocks).

Keywords: Chemical weathering; Japan; Lithology; Phosphorus; Sulfur; Calcium』

1. Introduction
2. Data and methods
　2.1. Overview
　2.2. Handling of geodata
　2.3. Mass balance
　2.4. Modeling technique
　2.5. Calculation of Ca-excess fluxes
　2.6. Application of the derived models to the Japanese Archipelago
3. Results and discussion
　3.1. Overview
　3.2. Calculated CSRWR based on the monitoring data
　3.3. Correlations of CSRWR
　3.4. Predicted CSRWR by the applied models
　　3.4.1. Overall interpretation for the monitored catchments
　　3.4.2. Average CSRWR predictions per lithological class
　　3.4.3. Physical interpretation of the b-estimates of the models
　　3.4.4. Residual analysis
　3.5. Application to the Japanese Archipelago
　3.6. The influence of excess-Ca from non-silicate calcic minerals on CSRWR of silicate-dominated lithological classes
　3.7. Contribution of sulfur to CSRWR&S
　3.8. Liberation of lithogenic phosphorus by chemical weathering
　3.9. Application to the global scale
4. Conclusion
Acknowledgements
Appendix A
Appendix B
Appendix C
Appendix D
Appendix E
References