Monitoring the dynamics of surface water fraction from MODIS time series in a Mediterranean environment

Development and Validation of a Dense 18-Year Time Series of Surface Water Fraction Estimates from MODIS for the Mediterranean Region

“…The prediction accuracy of our model (R 2 = 0.91, RMSE = 11.41%, MAE = 6.39%) is higher than the R 2 of 0.625 reported by Weiss and Crabtree (2011) who used a linear regression model, and R 2 of 0.7 reported by Guerschmann et al (2011) using a logistic regression model. This research successfully expanded our previous work (Li et al, 2018) by upscaling it from a relatively small region to the whole Mediterranean while retaining a similar high accuracy (both achieved an R 2 of 0.91). This attributes to the feasibility and 10 robustness of Cubist regression model to deal with different environmental conditions when training data are collected across a wide geography resulting in varying spectral characteristics.…”

“…The approach used to derived surface water fraction builds on our previous work (Li et al, 2018) with considerable improvements regarding input data, training data and commission error processing. We explored the use of MODIS spectral information and a topographic metric for estimating surface water fraction over two study areas in Spain through the use of rule-based regression models and concluded that a single global regression model can be effectively tuned locally as long as it is fed with training data that comprise the various environmental conditions encountered across the larger area (Li et al, 2018). In this sense, the approach for constructing a global model can be expanded effectively to wider areas such as the Mediterranean region.…”

“…To overcome this limitation and incorporate small water bodies, several researchers have attempted to predict sub-pixel surface water estimates of MODIS by providing water fraction in each pixel using techniques like linear spectral mixture modeling (e.g. Li et al, 2013;Olthof et al, 2015;Hope et al, 1999) and machine learning (e.g., Rover et al, 2010;Sun et al, 2012;Li et al, 2018) for small areas. However, the utility and efficiency of these methods have rarely been explored for the estimation of surface water fraction for larger areas.…”

“…However, the utility and efficiency of these methods have rarely been explored for the estimation of surface water fraction for larger areas. In our previous work (Li et al, 2018), we explored the use of rule-based regression models over two small areas on the Iberian peninsula and concluded that a single global regression model can provide accurate surface water estimates across areas with different environmental conditions as long as it is fed with training data that comprise these various conditions. Consequently, we concluded that this approach has potential to be applied over much larger areas.…”

A new dense 18-year time series of surface water fraction estimates from MODIS for the Mediterranean region

“…To overcome this limitation and incorporate small water bodies, several researchers have attempted to predict subpixel surface water estimates of MODIS by providing the water fraction in each pixel using techniques like linear spectral mixture modeling (e.g., Hope et al, 1999;Li et al, 2013;Olthof et al, 2015) and machine learning (e.g., Li et al, 2018;Rover et al, 2010;Sun et al, 2012) for small areas. However, the utility and efficiency of these methods have rarely been explored for the estimation of the surface water fraction for larger areas.…”