This work presents a model that is capable of predicting fuel characteristics. Fuel load and moisture is divided into live and dead fuels, and also includes wood and foliage components. One of the main outcomes of this work is a dataset of these variables on a daily scale and at ~9 kilometer spatial resolution. Overall, I found this work to be original and a much-needed contribution to the science.

However, I believe the manuscript may be improved if the following changes are made:

Lines 130-140: Didn’t understand what SW_live represented fully. May you please elaborate on what “the remaining component of live wood” means?

Line 156: Should this be +4.6 Pg yr-1 based on the numbers presented?

Section 2.2: Is it possible to add a table that summarizes everything outlined here? While it is well-written, I feel that the density of information is quite high and may benefit through using a table to organize all the numbers and where they come from.

Figure 6: Great figure! I would suggest placing the text at the top of each panel so the text is not in the way of the figure.

Figure 8: Same comment as Figure 6 (text is a bit hard to read in the bottom four panels).

Figure 9: Same comment as Figure 6 (text is a bit hard to read in the bottom two panels).

Figure 10: Same comment as Figure 6.

Figure 11: Same comment as Figure 6.

Figure 12: Would it be possible to increase the font size of the names and numbers along the x-axis and y-axis? Also for the legends? I found these figure panels a bit difficult to read due to small size.

Figure 13: Same comment as Figure 12.   

I have also added some of my edits/comments as a track changes document in the supplement file. 

After these changes are made, I believe this manuscript may be accepted for publication.

McNorton and Di Giuseppe present a new global model of biomass / fuel load and fuel moisture to aid efforts to better understand variability in fire activity and better predict wildfires. The approach is rooted in ESA-CCI biomass data but combined with other datasets to get temporal variability. To move from standing tree biomass to fuel loads (including litter) the authors used ratios of dead to live biomass based on the literature. Also satellite data of leaf area index are used as well as quite a bit of parameterization. Then the fuel moisture content of these different fuel classes is modeled.

The paper is well written and the methods are clear. I see the need for this work but have two major concerns that need to be addressed before publication

1) There is a very strong focus on standing tree biomass (both from a methodological point of view and for evaluation). Clearly this is important but standing live biomass is often not the main fuel source for fires. For example, in L147 the authors state that fuel loads in the Boreal region in the summer are 10% dead fuels. In general, however, emissions there stem for the vast majority from dead fuels (organic soil) according to for example the ABoVE campaign (Walker et al., 2020). Also in many other biomes the surface fuels area key, and models that aim to say something about fire danger should therefore (also) focus on surface fuels. In the current paper these are modeled, but seem of secondary importance and most of the evaluations are on standing biomass. One potential way forward is to evaluate the new dataset with the data from Walker et al. (2020) and the literature review by Van Wees et al. (2022) which specifically focuses on those papers that studied biomass from a fire perspective. Somewhat related, I was also wondering how realistic the large (sometimes doubling) seasonal changes in live and dead wood are (Figure 2 and 3)?

2) It is good to see that the soil moisture values are calibrated / compared to in situ data. The correlation is rather poor though with on average about 25% of the variability being explained (even lower for agriculture but the authors provide a good reason for that). I fully realize a perfect fit will be impossible but I respectfully doubt how useful the model is in this case. Simple example (L464): “Seasonal fire activity is reasonably well captured by both FSI (R = 0.58) and DFMC (R = 0.38).” In most fields of research these are low to moderate correlations; a model that in a range of evaluations shows little correlation may not be fit for the purpose. Clearly the evaluations of total AGB are more promising but as mentioned in 1) they may be less relevant. One way forward would be to iteratively adjust the parameterisations (not just for the soil moisture but in all steps) until the best comparison with evaluation data is found in some optimisation exercise. If this exercise shows that much variability is still not captured the authors need to re-think their approach

References:

Walker, X. J., Rogers, B. M., Veraverbeke, S., Johnstone, J. F., Baltzer, J. L., Barrett, K., Bourgeau-Chavez, L., Day, N. J., de Groot, W. J., Dieleman, C. M., Goetz, S., Hoy, E., Jenk- ins, L. K., Kane, E. S., Parisien, M.-A., Potter, S., Schuur, E. A. G., Turetsky, M., Whitman, E., and Mack, M. C.: Fuel availability not fire weather controls boreal wildfire sever- ity and carbon emissions, Nat. Clim. Chang., 10, 1130–1136, https://doi.org/10.1038/s41558-020-00920-8, 2020

van Wees, D., van der Werf, G. R., Randerson, J. T., Rogers, B. M., Chen, Y., Veraverbeke, S., Giglio, L., and Morton, D. C.: Global biomass burning fuel consumption and emissions at 500 m spatial resolution based on the Global Fire Emissions Database (GFED), Geosci. Model Dev., 15, 8411–8437, https://doi.org/10.5194/gmd-15-8411-2022, 2022.

(the database is mentioned under ‘Code and Availability’, direct link: https://doi.org/10.5281/zenodo.7229039)