Ocean alkalinity enhancement (OAE) can potentially remove gigatons of CO₂ from the atmosphere. By adding alkaline substances into seawater, the enhanced alkalinity will yield net removal of atmospheric CO2 if this enhancement happens at the surface. Consequently, it has long been assumed that these substances must be ground into small particles or pre-dissolved to prevent from sinking out of the surface before the addition to the open ocean. However, given the unique hydrodynamics of coastal systems, the applicability of this assumption remains to be verified. In this study, we developed a regional high-resolution model to simulate the addition, dissolution, and sinking of various alkaline feedstocks, and to evaluate their impacts on detectability, potential risks, and CO2 removal efficiency of OAE. Results show that in winter, a fast sinking velocity of feedstocks greatly reduces potential risks, while having minimal impact on the detectability of OAE signals or removal efficiency due to strong vertical mixing. In contrast, summer, which features strong stratification and long residence time, allows a slow dissolution rate to mitigates potential risks without sacrifice on detectability or efficiency. Thus, the potential risks can be mitigated without compromising detectability and efficiency by tailoring feedstock selection to local hydrodynamic conditions.