Surge hazards induced by tropical cyclones have caused substantial economic losses and casualties for coastal communities worldwide. Under projected sea level rise (SLR), it is not well understood how the probabilistic surge hazard may change. Coastal planning and engineering usually adopt the "bathtub" method to evaluate the surge-SLR response, where surge with SLR is considered to be the exact summation of the two. This method has been shown by previous studies to be unreliable. Studies on surge-SLR response either use a low-fidelity model setup or rely on individual storm's surge to represent the probabilistic surge due to the high computational burden. Herein, we use high-fidelity numerical models in the Tampa region, West Florida, consider 188 synthetic storms and four SLR scenarios, and investigate the surge-SLR response and its physical drivers. Compared to the direct summation of present-day surge and SLR amount, results show that the probabilistic surge with SLR can be 1.0 m larger, while different individual storm's surge with the same magnitude can be 1.5 m larger or 0.1 m smaller, indicating the importance of not relying on a limited number of surge events to assess the probabilistic surge response to SLR. Investigation of the physical drivers shows that distinct topographic features of the study area and storm forward speed notably affect the surge-SLR response. When considering 1.3 m or larger SLR in the study area, complex topography, and large surge events, the effects of SLR on the probabilistic surge are hard to predict and should be investigated more carefully. Plain Language SummaryCoastal flooding during hurricanes have caused substantial economic losses and deaths worldwide. The statistical flooding, for example, flooding with a 100-year return period, is an important concept for engineers and planners to build safe coastal communities. Under predicted sea level rise (SLR), how the statistical flooding will change is unclear. Simply adding the amount of SLR on top of current flooding value to get the future statistical flooding is a popular assumption but has been shown by to be unreliable. More recent studies use computer simulations to get more accurate results. However, this simulation takes a long time. Thus, researchers either simulate 5-10 selected hurricanes or simplify the physics. Here we use high-accuracy computer models in the Tampa Bay region, West Florida; consider nearly 200 storms and four SLR scenarios; and investigate how the statistical flooding will change due to SLR, as well as why. Results indicate the importance of not relying on a limited number of storms to assess the statistical flooding change due to SLR. When considering 1.3 m or larger SLR in the study area, complex topography, and severe flooding events, the effects of SLR on statistical flooding are hard to predict and should be investigated more carefully.Projected sea level rise (SLR) can change both deterministic and probabilistic surge responses, causing uncertainties for coastal planning and engineering. Fo...

• Physical scaling laws are revealed to characterize forerunner surge from storm track parameters (central pressure, radius, and speed). • The physical scaling enables rapid forecasting of forerunner surge.