The last few years have seen a dramatic increase in the amount of data available to model Earth and environmental systems thanks to new sensing technologies and open data policies. To make the best use of these newly available data, innovative modeling approaches are being developed, such as machine learning, which is ideal to extract valuable information from large amounts of data. This applies particularly well to modeling the water cycle, where non-linear processes are ubiquitous. For example, complexities occur at the large scale, e.g. when considering global feedbacks between hydroclimatic variables, or at smaller scales, e.g. when modeling the hydrological response of a catchment after a storm, or the spread of a contaminant in an aquifer.
This Article Collection welcomes submissions that focus on the application of machine learning approaches to better predict and understand water resources behaviors. We envision papers that have, at the same time, a component of machine learning algorithms (e.g. applications or developments in neural networks approaches in their various forms, non-parametric classification or regression on large datasets, non-parametric spatial processes modeling) as well as a component of water resources modeling (from global to local scale, encompassing e.g. atmospheric processes, surface hydrology, or subsurface flow and transport processes).
We particularly encourage in the following domains, although other topics might be of interest as well:
- Improved predictions of hydrological, hydrogeological or hydroclimatological variables
- New ways of using machine learning approaches to unravel hydrological processes (opening the black box)
- Application of machine learning in fields where it was not considered before
- Approaches where statistical learning can be seen as an advantageous alternative to physical description of a hydrological system
- Ways to address scale dependencies between punctual and areal measurements of the water cycle
The last few years have seen a dramatic increase in the amount of data available to model Earth and environmental systems thanks to new sensing technologies and open data policies. To make the best use of these newly available data, innovative modeling approaches are being developed, such as machine learning, which is ideal to extract valuable information from large amounts of data. This applies particularly well to modeling the water cycle, where non-linear processes are ubiquitous. For example, complexities occur at the large scale, e.g. when considering global feedbacks between hydroclimatic variables, or at smaller scales, e.g. when modeling the hydrological response of a catchment after a storm, or the spread of a contaminant in an aquifer.
This Article Collection welcomes submissions that focus on the application of machine learning approaches to better predict and understand water resources behaviors. We envision papers that have, at the same time, a component of machine learning algorithms (e.g. applications or developments in neural networks approaches in their various forms, non-parametric classification or regression on large datasets, non-parametric spatial processes modeling) as well as a component of water resources modeling (from global to local scale, encompassing e.g. atmospheric processes, surface hydrology, or subsurface flow and transport processes).
We particularly encourage in the following domains, although other topics might be of interest as well:
- Improved predictions of hydrological, hydrogeological or hydroclimatological variables
- New ways of using machine learning approaches to unravel hydrological processes (opening the black box)
- Application of machine learning in fields where it was not considered before
- Approaches where statistical learning can be seen as an advantageous alternative to physical description of a hydrological system
- Ways to address scale dependencies between punctual and areal measurements of the water cycle