This study introduced a newly developed offline model of lateral N transfers, called ORCHIDEE-NLAT, within the framework of the land surface model ORCHIDEE. The ORCHIDEE-NLAT was used to simulate historical changes in riverine DON, PON, and DIN exports across the globe. Overall, it is an important work of global riverine N transport model development. The manuscript is well written and the model structure is clearly illustrated. Currently, the accuracy of the model in simulating riverine N exports is actually low, especially at regional scale. I understand it is very challenging to accurately simulate N transfers at the global level, but I still have some suggestions for authors to improve the model in the future.

1. The ORCHIDEE-CNP and ORCHIDEE-Clateral are both used to provide the land-to-river inputs. ORCHIDEE-Clateral provides runoff, drainage, DOC, and POC inputs, while ORCHIDEE-CNP provides inputs of DON and DIN leaching from manure. Can ORCHIDEE-CNP provide all these forcing data? Using the outputs of two models may bring uncertainties and make this study complicated. Since runoff and drainage are critical components that determine DIN, DON, and PON fluxes, different water inputs simulated by two versions of land models can bring inconsistencies in water flux information behind N fluxes.
2. In the aquatic N module, why not consider the transformation process from PON to DON, and from inorganic N to organic N?
3. The residence time method is used to calculate N transport along the river networks. This method is commonly used but very simple and may not be able to accurately capture water transport processes. Authors may consider using hydrological kinetic equations in the future.
4. The validation of model results only focuses on TN and NO3. How to validate DON and PON flexes? USGS provides organic N observation. The assumption of a linear relationship between observed TN and NO3 may ignore the variations in organic N.
5. The simulated total N fluxes in the 1920s is questionable as authors have already mentioned. What are climate data sources? How about the precipitation change? ORCHIDEE has already been used to simulate lateral C and sediment fluxes, and does the same issue occur in these simulated variables? Better to check it and make it right.

Minor

L64. N leaching into the aquatic environment. LOAC includes land.

L91-93. This is also an issue for LSM.

L157-159. Does ORCHIDEE-CNP have soil organic C pools? It should also have POC outputs.

L166-167. How to separate sewage TN into different N species.

L197. A constant ratio may make the simulated DON less informative and accurate.

L257. What about N deposition into sediment?

Equations 4-11. Not all variables are explained. Please check.

Figure 4. Put the NSE value into the figure.

L507. Can you explain the decrease in PON export?

Figure 7. Better to use mg/L as concentration unit.

L586. Figs.

L625-627. ON inflow is simulated by ORCHIDEE CNP and Clateral, not NLAT. Right?

Figure 11c. Wrong name of NLAT.

L738-741. Another important thing is to improve model structure and data quality.

L863. Doi is invalid.

see attached PDF