A Larval “recruitment Kernel” to Predict Hatching Locations and Quantify Recruitment Patterns

Larval recruitment, a critical component of population connectivity, has been under investigated compared to larval dispersal. We developed a backward-in-time Lagrangian particle tracking model to predict larval hatching locations and proposed a larval recruitment kernel, to quantify recruitment patterns. Combining field data and a hydrodynamic model, our backtracking model predicted Lake Whitefish (Coregonus clupeaformis) hatching locations in Lake Erie. We found a strong linear correlation (r = 0.95-0.98) between travel distance (i.e., distance along a trajectory) and pelagic larval duration (PLD), and a moderate correlation (r = 0.66-0.68) between linear distance (i.e., displacement) and PLD. This questions the wide use of PLD as a proxy for dispersal distance. We defined the recruitment kernel using the probability density function of the linear recruitment distance. Characteristics of the recruitment kernel, such as theoretical self-recruitment, median-recruitment distance, long-distance recruitment, and openness convey significant information about population connectivity that are distinct from those derived using the well-known dispersal kernel (e.g., theoretical local retention). The dispersal kernel has been widely and successfully applied to quantify dispersal patterns of plant seeds, insects and fish larvae. Due to the complexity of in situ observations tracking small-sized larvae, forward-in-time Lagrangian particle tracking models have been widely applied to predict and quantify larval dispersal patterns, by releasing particles from the spawning/hatching region and estimating the dispersal kernel (i.e., the probability density function, p.d.f., of linear dispersal distance of particles). Whereas it remains a challenge to predict and quantify larval recruitment, which is another critical component of population connectivity, by releasing particles from every potential hatching region, and discriminating every recruit at the settlement sites. A backward-in-time particle tracking model was applied here to predict larval hatching locations by releasing particles at larval nursery locations. Based on the p.d.f. of the linear recruitment distance, a larval recruitment kernel was first proposed to quantify recruitment patterns in the case where spawning/hatching sources are unknown. The proposed recruitment kernel holds distinct ecological significance compared to the dispersal kernel. Characteristics of the recruitment kernel, for example, the self-recruitment, is distinct from those of the dispersal kernel, for example, the well-known local retention, providing important supplements to population connectivity research. Larval Whitefish sampled in Lake Erie's western basin were backtracked to hatch in the western basin, while not around the Bass Islands A larval recruitment kernel was proposed to quantify larval recruitment patterns and allow for a theoretical measure of self-recruitment Moderate correlation between pelagic larval duration (PLD) and linear distance questions the use of PLD as a proxy for dispersal potential