@article {bnh-8100, title = {Continental-scale prediction of live fuel moisture content using soil moisture information}, journal = {Agricultural and Forest Meteorology}, volume = {307}, year = {2021}, month = {06/2021}, abstract = {

Live fuel moisture content (LFMC) is a key factor that determines the flammability of vegetation in ecosystems. Soil moisture (SM) is one of the variables that is known to influence plant water use. The present study analyses the LFMC-SM relationship over Australia using gridded, remote sensing-based LFMC and land surface model-based SM products. A lag-correlation analysis conducted over 60 selected sites shows that the strength of the relationship between LFMC and SM varies from site to site and, in general, is moderately strong (median lag-correlation of ~0.5). However, the strength of the relationship changes with vegetation type and also with soil profile depth. At all the sites, SM is found to be a leading indicator of LFMC. The lag also varies with the location and is found to range from days to months. Based on the location-based correlation analysis, we identify the 0-35 cm SM profile (SM0-35cm) to be the best predictor of LFMC. We developed a simple model to predict daily LFMC, where it is hypothesised that daily variations in LFMC from its annual cycle can be predicted using daily deviations from the annual cycle in SM0-35cm. The annual cycles of LFMC and SM0-35cm are modelled using Fourier cosine series. The averaged (over 60 sites) correlation obtained for the validation period is 0.74 when a time-lag of 14 days is assumed at all locations. When the model is applied nationally at a 5 km grid, the normalised root mean squared error for the validation period is found to be less than 25\% in general. The results from the present study highlight a modelling strategy that can be used to address a critical gap in the forecast of spatially and temporally continuous LFMC at regional scales in advance for operational fire management applications.

}, keywords = {AFMS, Fire, JASMIN, JULES, live fuel moisture, Soil moisture}, doi = {https://doi.org/10.1016/j.agrformet.2021.108503}, url = {https://www.sciencedirect.com/science/article/abs/pii/S0168192321001866}, author = {Vinod Kumar and Imtiaz Dharssi and Marta Yebra and Paul Fox-Hughes} } @article {bnh-6237, title = {Disaggregation of JASMIN soil moisture product to 1KM resolution}, number = {521}, year = {2019}, month = {12/2019}, institution = {Bushfire and Natural Hazards CRC}, address = {Melbourne}, abstract = {

Fire intensity, spread rate, and ignition are very sensitive to fuel dryness which in turn is strongly linked to soil moisture deficit. Though the value of soil moisture deficit in predicting fire danger has been long established, very few fire danger rating systems employ a comprehensive methodology to estimate it. Most such fire danger rating systems use very simple empirical water balance models which are found to have errors. The Bureau of Meteorology has recently developed a prototype, highresolution, land surface modelling based, state-of-the-art soil moisture analysis for Australia. The product, called Joint United Kingdom Land Environment Simulator (JULES) based Australian Soil Moisture Information (JASMIN), has a spatial resolution of 5 km at hourly timesteps. However, applications like fire danger mapping may require soil moisture information at higher spatial resolution due to the large spatial variability of soil moisture in the landscape. We focus on some of the research carried out to downscale the JASMIN product from 5 km to 1 km spatial resolution. We discuss the application of three downscaling algorithms: two regression-based methods and one with a theoretical basis. The three methods applied are based on the well-known surface temperature {\textendash} vegetation index space. We present an overview of the application of each method, along with an evaluation and comparison against each other and against ground-based soil moisture observations. Results from comparison with ground-based soil moisture measurements indicate that there is no significant degradation of the bias in the three methods, when going into higher spatial resolutions. However, the regression methods, in general, fail to capture the observed temporal variability. The theoretical based method, on the other hand, provides a temporal correlation of 0.81 and captures the skill of the parent JASMIN product.

}, keywords = {dryness, Emergency management, JASMIN, resolution, Soil moisture}, issn = {521}, author = {Vinod Kumar and Imtiaz Dharssi and Paul Fox-Hughes} }