In the first Monte Carlo (MC) run collection, the merged rain estimates had a factor range of 0.2 to 1.35 (see factor 24 in Table 1), reflecting the radar depth bias across the study events (Table 2). Figure 7a shows the all-event influence of the rain depth bias factor (RainM; refer to Table 1). The total-order indices show the dominant influence of RainM, highlighting the importance of depth bias-free rain inputs for accurate forecasts. LFB has a second-order effect, though most of the individual peaks have negligible second- and first-order effects (not shown). This would have made it difficult to detect the rain influence from the behavioral classification using sensitivity analysis (SA) methods which do not explicitly consider interaction at all orders (e.g., RSA or regional SA with a second-order factor correlation analysis). The rain influence using RSA is probably better detected from LFs like the multiobjective LFM on which RainM is seen to have a significant first-order effect. Also, use of identifiability plots may help to constrain the RainM uncertainty in regions without rain gauge measurements, potentially allowing the universal application of radar rain estimates under the current large depth biases