Imagine you're body-surfing. A wave that hits you at a glancing angle shoves you sideways — annoying but manageable. A wave that hits you square in the chest knocks you flat. That's the difference between an oblique and a perpendicular tropical cyclone landfall, except the 'you' is a coastline and the knockdown force is storm surge. This paper asks: is climate change turning more of those glancing blows into direct hits? The committed claim: under high-emissions warming (SSP5-8.5), the global fraction of near-perpendicular tropical cyclone landfalls increases significantly through 2099, with the strongest signal along the eastern North Pacific. Under intermediate emissions (SSP2-4.5), no such global trend emerges. This makes landfall angle an emissions-pathway-dependent dimension of coastal risk that nobody was systematically tracking. The method is PepC-Global, a statistical-dynamical tropical cyclone model that generates large synthetic storm sets driven by environmental fields — either from reanalyses (for validation) or from 14 CMIP6 general circulation models (for projections). The model doesn't just simulate storm tracks; it resolves approach angles near landfall with enough fidelity to classify perpendicular vs. oblique hits. The validation step is key: the authors show that PepC-Global reproduces observed landfall-angle distributions and the geographic pattern of near-perpendicular fractions from 1980-2024 IBTrACS data. The simulations also reproduce the fact that the 1980-2024 observed trend is weak and statistically insignificant — a reassuring negative result that builds confidence before the projections go forward. On the ladder: there is essentially no prior art that directly addresses landfall angle under climate change. Studies of TC track changes, translation speed, and intensity abound (Kossin 2018 on slowdown, Knutson et al. 2020 on intensity), but nobody had systematically decomposed the landfall-angle distribution under warming. The paper creates a new rung rather than climbing an existing one. The 14-model CMIP6 ensemble is the current gold standard for forced projections, though any statistical-dynamical downscaling inherits the GCM biases in large-scale steering flow. The integrity story is solid for this subfield. The validation against IBTrACS is honest: 1980-2024 trends are insignificant in both observations and simulations, which means the model isn't hallucinating a signal where none exists. The 14-model spread gives a real uncertainty range. The separation between SSP2-4.5 (no significant trend) and SSP5-8.5 (significant trend) is a natural internal control — if both scenarios showed the same signal, you'd suspect model artifact. The main caveat is that PepC-Global is one modeling framework; independent reproduction with a different synthetic TC generator would substantially strengthen the result. The mechanism behind the perpendicular-fraction increase is important: it's not that storms are simply moving to different coastlines. The increase arises from changes in near-perpendicular fractions at individual coastal locations — meaning the atmospheric steering patterns that guide final approach are themselves shifting. This is a subtler and more physically grounded signal than a simple redistribution story. The practical implication is pointed. Storm-surge risk assessments, coastal engineering standards, and flood insurance models that use historical landfall-angle distributions may be underestimating future surge hazard under high emissions. A 10-15% increase in perpendicular-fraction doesn't sound dramatic until you remember that storm surge scales nonlinearly with the component of translation velocity normal to the coast. The paper identifies the eastern North Pacific as the hotspot — Mexico's Pacific coast and parts of Central America face the most spatially coherent increase.
