Why distance from the road is considered a potential input feature in machine learning modeling of Earth science problems?
In recent years, researchers have started incorporating distance from the road-rail network as an input feature in solving Earth science problems through machine learning. Have you ever think how this feature is going to impact the response or target variable? In this article, we will try to understand this question by using a suitable case study that has been recently published. You can download the paper for more details. You can write to me (abhilash.singh@ieee.org) if you have any questions or visit my web page for more updates.
Anthropogenic activities often cause the problem of drainage congestion. The development of road-rail network act as a anthropogenic barriers which lead to the accumulation of the soil moisture near these physical barriers. In a recent study, Singh et al. 2022, used soil moisture as a proxy to study and quantify the impact region of the drainage congestion due to road network. In doing so, they have first use Sentinel-1/2 satellite imagery to estimate high resolution surface soil moisture map and then they studied the spatial pattern of soil moisture near the road network. Interestingly they found that irrespective of the road orientation (i.e., vertical, horizontal, and inclined; Figure 1) the soil moisture in the proximity of the road-network is very high which decreases gradually as we move away from the road (Figures 2–4).
Based upon this study, the conclude the following points;
- Soil moisture information can be used as a precursor to measure the drainage congestion along the road network.
- Road network acts as a physical barrier and leads to drainage congestion at several places. This eventually results in high soil moisture in the proximity of the road network.
- The extent of drainage congestion is different for the road network oriented in vertical, horizontal, and inclined direction. The extent of drainage congestion is relatively more at the locations where roads traverse in inclined direction.
Through this study we get to know that the soil moisture is high near road network. We also know that soil moisture plays a very crucial role in the regional climate change and has a wide application in climate change, agronomy, water resources, and in many other domain of science and engineering. Hence either directly or indirectly, soil moisture is influencing each segment related to the earth science. So, the road network is going to influence the pattern of soil moisture which ultimately controls the target variable or response variable in the earth science machine learning problems.
References
- Singh, A., Naik, M. N., & Gaurav, K. (2022). Drainage congestion due to road network on the Kosi alluvial Fan, Himalayan Foreland. International Journal of Applied Earth Observation and Geoinformation, 112, 102892.
- Singh, A., Gaurav, K., Meena, G. K., & Kumar, S. (2020). Estimation of soil moisture applying modified dubois model to Sentinel-1; a regional study from central India. Remote Sensing, 12(14), 2266.
