Dual crop coefficient approach of partitioning evapotranspiration (ETc) into transpiration (T) and evaporation (E) has been used extensively for applications ranging from estimating crop water allocations to irrigation scheduling. Although the basal crop coefficient (K-cb) provides an improved estimate of T, variations in K-cb are still possible due to variations in climate and management practices necessitating the development of regional K-cb. Large drainage lysimeters (4.87 m x 3.65 m x 1.37 m) were used to develop K-cb for drip irrigated bell pepper and watermelon with plastic mulch in sub-tropical Florida using four and three seasons of data, respectively. The average K-cb values for the initial, mid-season, and late stages were 0.12, 0.68, and 0.77, respectively, for bell pepper, and 0.05, 0.96, and 0.66 for watermelon. The K-cb values for pepper from this study were statistically lower (p= 0.047) than the generic FAO-56 values (adjusted for mulch and plant density) and improved the estimates of T and E by 27.3 and 7%, respectively. Although lysimeter K-cb values for watermelon were numerically lower than FAO-56, no statistical difference was detected. However, FAO-56 overestimated watermelon E by 52%. The lysimeter-based K-cb improved the ETc estimate and this improvement was 26 and 51% of the respective seasonal rainfall for pepper and watermelon. When extrapolated to all drip irrigated pepper acreage in Florida, the K-cb from this study could potentially reduce the applied water by some 6.9 million m(3) of water compared to using FAO-56 K-cb, highlighting the importance of accurate estimation of ETc for the irrigation management. The improved K-cb will help customize the irrigation management and reduce nutrient leaching as well as improve simulations of ETc within the hydrologic models for similar environment (C) 2014 Elsevier B.V. All rights reserved.