DYNAMICS OF ABOVEGROUND NET PRIMARY PRODUCTION IN A DESERT GRASSLAND OF ARIZONA

Abstract

Studies were conducted on the dynamics of live, standing recent dead, and standing old dead herbage for a desert grassland site in the southwestern United States. Data were collected at eleven sample dates from August 24, 1979 to October 18, 1980. Coefficients of variation for biomass data for most species exceeded 400% with only the most abundant species sampled with coefficients of variation less than 100%. Coefficients of variation generally were less than 80% for the perennial grasses as a group. The mean rate of transfer of biomass from live herbage to standing recent dead herbage was 0.21% per day for 203 days when expressed as a percentage of the peak live herbage. A rate of transfer of 0.53% per day of live herbage to standing dead for 109 days for perennial grasses in the desert grassland was found during early vegetative growth. A high proportion of leaves to culms was present during this period as compared to the summer growing season. Heavy precipitation in the early part of September accelerated growth of live herbage, and also accounted for losses of standing recent dead and standing old dead herbage to the surface litter component. The loss of dry leaves from standing grass culms due to rainfall impact contributes to an underestimate of net primary production. Standing old dead biomass accumulated following the depression in September. Transformation of the recent dead herbage to an appearance of old dead late in the summer also contributed to underestimating of current year production based on peak crop method due to the identification of recent dead herbage as the old dead component. Summation of growth increments by sample periods provided the highest estimate of aboveground net primary productivity when compared to estimates based on peak standing crop, summation of species peaks, and summation of significant growth increments by sample periods. Estimates of net productivity were 186% of the estimate at peak standing crop. Estimates of net production based on significant growth increments by sample periods yield the lowest value for net production, because poor precision of samples restricted data to few significant values. Based on the peak periods for standing biomass of lifeform groups, a minimum of four sample dates are recommended to estimate aboveground net primary production for semidesert grassland in southeastern Arizona. These sample dates include: (1) late April to sample spring peaks by perennial grasses, annual forbs and annual grasses; (2) late June to capture the spring trough; (3) late September to sample peak crops of perennial forbs, and half-shrubs; and (4) mid to late October to sample the fall peak for perennial grasses. Net production estimates by the peak crop method for perennial grasses varied from 18.7 g/m² in 1979 to 86.1 g/m² in 1980. This variability was attributed to the differences in the amount of rainfall in 1979 and 1980. This factor, therefore, largely influenced the repeatability of biomass estimates from one year to another. The large amount of time (approximately 644 man-hours) needed to sample and separate live and dead components by species for 20 quadrats at each harvest date, together with the year-to-year variability and the need for a minimum of four sample dates to effectively sample the desert grassland community, makes biomass data impractical for general use as a data base in land-use planning. However, this effort should be looked at from the standpoint of research studies describing the dynamics of range systems based on temporal characteristics and transfer functions to provide data which can help design management strategies for these range ecosystems.