The study investigates predictions of surface air temperature under strong and weak L-A coupling conditions using NASA’s GEOS S2S forecast system. The authors assess the predictive skill of temperature forecasts during weeks 3 and 4 of the boreal summer, particularly in the Midwest and northern Great Plains, by applying L-A coupling metrics. These metrics capture the relationships between soil moisture, latent heat flux, and some other atmospheric variables, such as surface skin temperature and planetary boundary layer height. The study hypothesizes that surface temperature predictions improve when strong coupling is present. The results show that strong L-A coupling increases the ability to predict extreme warm events, especially for the northern Great Plains, where strong coupling increases the likelihood of correctly predicting abnormally warm temperatures during weeks 3-4.

While the paper is well-written and concise, the discussions are limited and could be expanded to address several key issues. For example, it should be addressed how some preprocessing explained in the method section, such as upscaling or spatiotemporal aggregation of the different datasets to match each other, plays a role in the results. Moreover, the strength of land-atmosphere coupling is mostly mediated by soil moisture, especially in water-limited regions. Despite its importance, there are no discussions about the accuracy of the soil moisture used in this study. Even a small bias in soil moisture values used in this study, especially in heavily irrigated regions during the growing season, may have a significant impact on the subseasonal air temperature predictions. Providing a comparison with observational-based soil moisture observations such as SMAP would more clearly identify the regions where the strength of land-atmosphere coupling is more reliable in contributing to air temperature prediction skills.

Dear Authors,

Please see my comments in the attached file for detailed feedback.

Best regards,

Husain Najafi