Abstract By mediating evapotranspiration processes, plant canopies play an important role in the terrestrial water cycle and regional climate. Substantial uncertainties exist in modeling canopy water interception and related hydrological processes due to rainfall forcing frequency selection and varying canopy traits. Here we design a new time interpolation method “zero” to better represent convective‐type precipitation in tropical regions. We also implement and recalibrate plant functional type‐specific interception parameters for rainforests and oil palm plantations, where oil palms express higher water interception capacity than forests, using the Community Land Model (CLM) versions 4.5 and 5.0 with CLM‐Palm embedded. Reconciling the interception scheme with realistic precipitation forcing produces more accurate canopy evaporation and transpiration for both plant functional types, which in turn improves simulated evapotranspiration and energy partitioning when benchmarked against observations from our study sites in Indonesia and an extensive literature review. Regional simulations for Sumatra and Kalimantan show that industrial oil palm plantations have 18–27% higher transpiration and 15–20% higher evapotranspiration than forests on an annual regional average basis across different ages or successional stages, even though the forests experience higher average precipitation according to reanalysis data. Our land‐only modeling results indicate that current oil palm plantations in Sumatra and Kalimantan use 15–20% more water (mean 220 mm or 20 Gt) per year compared to lowland rainforests of the same extent. The extra water use by oil palm reduces soil moisture and runoff that could affect ecosystem services such as productivity of staple crops and availability of drinking water in rural areas. , Plain Language Summary This manuscript addresses two issues in modeling canopy rainfall interception and associated water cycle processes in the Community Land Model and applies the model development to simulate a rainforest to oil palm transformation system. First, interpolating precipitation data from a coarser input frequency (e.g., three‐hourly or six‐hourly) to a finer model timescale (half‐hourly) with a nearest time method would introduce unrealistic rain events and bias the water cycle simulation for regions with convective‐type (short and intense) precipitation. Second, different canopy structures and leaf and stem morphologies, such as those of oil palm and rainforest, offer different water interception capacities, which questions the validity of a uniform interception parameterization across plant types used in land surface models. We design a new time interpolation method and implement plant type‐specific interception parameters in the Community Land Model, which produce more realistic precipitation events for tropical climate and predict more accurate canopy evaporation and transpiration for both oil palm and forest compared to observation. Results from our improved modeling approach indicate that current industrial oil palm plantations in Sumatra and Kalimantan (Indonesia) increase water use and reduce soil moisture and surface and subsurface runoffs, which could affect ecosystem services that rely on soil water reserve and river discharge. , Key Points We design a new time interpolation method “zero” to better represent convective‐type precipitation in tropical regions Using PFT‐specific interception parameters improves simulated partitioning of evapotranspiration for both oil palm and forest Current oil palm plantations in Sumatra and Kalimantan use 15–20% more water (211–238 mm) per year compared to lowland rainforests