Examining post-wildfire reseeding on arid rangeland: A multivariate tobit modelling approach

Despite fire cycles of increasing severity and frequency in the arid western U.S., very little empirical analysis has examined the success of plant seeding on arid rangeland following fire. This manuscript uses a unique dataset to assess causal factors underlying the measured densities, several years after fires, of: (1) unwanted invasive grasses, (2) seeded grasses, and (3) sagebrush, on rangeland in a western U.S. state. To accommodate various characteristics of the data, we employ trivariate tobit maximum likelihood estimation. Results indicate that the success of reseeding efforts is sensitive to the timing of and techniques used in reseeding, as well as whether grazing is allowed on the land. Our findings have broad implications for emergency fire rehabilitation management and policy for rangelands in the western U.S., as well as arid and semiarid rangelands elsewhere.