资源环境科技发展态势分析平台

The establishment and maintenance of an exhaustive hydrometeorological network are challenging tasks in densely populated coastal cities having erratic rainfall patterns. The current study proposes a robust statistical framework to rationalize an existing Automatic Weather Station (AWS) network which monitors multiple hydrometeorological observations to obtain maximum information at an optimal cost. This framework combines multivariate statistical approaches and multi attribute decision making techniques. We have demonstrated the proposed framework utilizing rainfall and relative humidity information from the existing AWS network across Mumbai city. Principal Component Analysis (PCA) is performed on daily rainfall and relative humidity datasets, followed by Technique of Order Preference by Similarity to Ideal Solution (TOPSIS) to rank the stations based on their capability to capture spatiotemporal variability. Out of the initial 60 AWSs established by the Municipal Corporation of Greater Mumbai (MCGM) with the primary objective of addressing Mumbai city's flood-susceptibility problem, 35 AWS data are used for this analysis on grounds of data completeness and reliability. Our analysis reveals that the spatiotemporal information of relevant hydrometeorological observations can be proficiently collected by a rationalized network of 22 AWS and that this may bring down network maintenance costs. The flood inundation and hazard maps for the Mithi catchment, one of the major flood hotspots of the city, are derived from the existing network and the rationalized network. The comparison of maps highlights the high accuracy of the rationalized network to reproduce the spatial characteristics of floods across the catchment. Our study presents a first-of-its-kind attempt to evaluate the performance of the rationalized network at flood inundation and hazard level derived from comprehensive 1-D and 2-D hydrodynamic approaches. The proposed generic framework can be utilized to reassess the precision and efficiency of various existing hydrometeorological monitoring networks at a regional-to-national scale and to achieve maximum spatiotemporal information from multiple hydrometeorological observations but at optimum maintenance cost.