Maintaining forage availability is challenging for managers of grazing systems, especially in spatially heterogeneous swards. Remote sensing may help to overcome this problem. The objectives of this study were to (i) determine a method by which NDVI may be calibrated to estimate biomass, (ii) determine if NDVI can be used to assess spatial variability of yield in extensive grasslands, and (iii) to determine if NDVI can be used to evaluate grazing systems. We found that the calibration of NDVI values for the estimation of biomass was better correlated with the destructive harvesting procedure (R2 = 0.68) but far more laborious and time-consuming than estimation of biomass from the rising plate meter (R2 = 0.54). Semivariograms revealed that sampling at a 0.76 m distance provided information about the spatial variability structure of NDVI values from grazed swards. Frequency distributions of sward biomass derived from NDVI reflected foraging strategies of cattle. Negative skewness and high kurtosis of histograms indicated selective grazing, while positive skewness and low kurtosis indicated the opposite. Histograms also allowed for estimation of available forage within each field. We concluded that grassland biomass may be derived from high resolution NDVI and RPM data and used to evaluate condition of grassland landscapes and aid decision-making of managed grazing systems.

A quantitative description of the foraging process is necessary for effective planning and execution of intensive grazing schemes. Foraging behavior is defined as having two components: feeding and moving. At intervals the foraging animal walks a number of steps searching for food and then pauses to feed at a new position here termed a feeding station. Five behavioral variables were analyzed under this framework: 1) time spent at a feed ing station; 2) number of bites at a feeding station; 3) steps taken between stations; 4) rate of steps; and 5) foraging time. The experimental design consisted of grazing small adjacent, approximately 7-ha paddocks for periods lasting 8 days. Animals significantly (P<0.01) increased the probability of taking 1 to 2 bites at a station as the season progressed. Regression analysis relating foraging time (in days) on a paddock revealed that the regression coefficients were statistically significant (P<0.05) suggesting that heifers were appreciably increasing foraging time as the grazing periods progressed. Analysis of moving behavior indicated that animals most often took 1 step between feeding stations and moved at approximately the same rate regardless of sward conditions. The significance of the behavioral measurements is discussed.