The study is timely and important in the context of the increasing pressures that climate change places on agriculture. It focuses on rainfed rice production in Nigeria and relates meteorological indicators (SPI and SPEI), as well as satellite-derived indicators (NDVI and soil moisture), to yield variability and drought occurrence reported at the farm level. 

The analysis is very detailed and there is no major issue with it. One minor concern is the mismatch in spatial scale between the meteorological indicators (25 km resolution) and the farm-level observations. However, this is not critical, as meteorological drought conditions are often considered spatially coherent. The explained variance is not particularly high, but this is expected when comparing meteorological indicators with farm-level data at such different spatial scales. However, the spatial resolution of the SPI and SPEI datasets is not reported and should be clearly stated in the manuscript as a potential limitation of the study. In any case, the improvement achieved by incorporating satellite-derived indicators is convincing, particularly during the drier year. This is the main result of the paper that deserves to be published.

As a suggestion for future work, the authors could also consider drought indicators derived from thermal remote sensing, such as ECOSTRESS and the Cooling Efficiency Factor Index (CEFI). These indicators have the potential to bridge the gap between driver-based drought indicators, such as SPI and SPEI, and impact-based indicators by directly detecting stomatal closure and the associated reduction in transpiration under drought stress. A brief mention of these approaches in the discussion could further strengthen the manuscript.   

The objective of this article is to evaluate the effectiveness of Earth Observation indicators (vegetation (NDVIA) and soil moisture (SWIA) anomalies) for monitoring the impact of droughts on rain-fed rice yields in Nigeria, in comparison with traditional meteorological indices (SPI, SPEI). The goal is to predict crop losses due to droughts and help farmers adapt.

To do this, the authors create various combinations of indices, covering different stages of rice growth, and calculate averages across areas of varying sizes.

Results show that models that include NDVIA and SWIA tends to better predict rice yields than those using only SPI/SPEI, especially for years when the meteorological indices fail (e.g., 2023) even if R² values are still relatively low.

The paper is clearly written and well-organized and the subject is appropriate to EGUSphere. I would regret the lack of rice yield timeseries in t/ha. There is no typical value of rice yield in the paper, only some RSME in t/ha in Figure A4. Nevertheless, I would recommend the paper for publication with only one minor comments.

Can authors add a figure showing rice-yield temporal evolution, and example of estimated rice yields using one or tow combination of indices.