Nariyasu Watanabe

Mapping herbage biomass and nitrogen status in an Italian ryegrass (Lolium multiflorum L.) field using a digital video camera with balloon system

Improving current precision nutrient management requires practical tools to aid the collection of site specific data. Recent technological developments in commercial digital video cameras and the miniaturization of systems on board low-altitude platforms offer cost effective, real time applications for efficient nutrient management. We tested the potential use of commercial digital video camera imagery acquired by a balloon system for mapping herbage biomass (BM), nitrogen (N) concentration, and herbage mass of N (Nmass) in an Italian ryegrass (Lolium multiflorum L.) meadow. The field measurements were made at the Setouchi Field Science Center, Hiroshima University, Japan on June 5 and 6, 2009. The field consists of two 1.0 ha Italian ryegrass meadows, which are located in an east-facing slope area (230 to 240 m above sea level). Plant samples were obtained at 20 sites in the field. A captive balloon was used for obtaining digital video data from a height of approximately 50 m (approximately 15 cm spatial resolution). We tested several statistical methods, including simple and multivariate regressions, using forage parameters (BM, N, and Nmass) and three visible color bands or color indices based on ratio vegetation index and normalized difference vegetation index. Of the various investigations, a multiple linear regression (MLR) model showed the best cross validated coefficients of determination (R2) and minimum root-mean-squared error (RMSECV) values between observed and predicted herbage BM (R2 = 0.56, RMSECV = 51.54), Nmass (R2 = 0.65, RMSECV = 0.93), and N concentration (R2 = 0.33, RMSECV = 0.24). Applying these MLR models on mosaic images, the spatial distributions of the herbage BM and N status within the Italian ryegrass field were successfully displayed at a high resolution. Such fine-scale maps showed higher values of BM and N status at the bottom area of the slope, with lower values at the top of the slope.

Distinguishing Cattle Foraging Activities Using an Accelerometry-Based Activity Monitor

Abstract Various sensors and analytic tools have been developed to assist with the collection and analysis of data regarding the activities of animals at pasture. We tested an accelerometry-based activity monitor, the Kenz Lifecorder EX (LCEX; Suzuken Co Ltd, Nagoya, Japan), to differentiate between foraging and other activities of beef cows in a steeply sloping pasture. Logistic regression (LR) and linear discriminant analysis (LDA), two of the most widely used techniques for distinguishing animal activities based on sensing device information, were employed in the analysis. An LCEX device was worn on a collar by each of four cattle over the course of 4 d, during which time the activity (foraging, resting, ruminating, walking, and grooming) of each cow was recorded by trained observers at 1-min intervals for a total of 15 h. LR and LDA were applied to the LCEX and observer data to distinguish between foraging and other activities. Overall, a more accurate measure was obtained by LDA (90.6% to 94.6% correct discrimination among cows) than by LR (80.8% to 91.8% correct discrimination). The threshold LCEX value for distinguishing between foraging and other activities varied among cows, and the correct discrimination rate for the pooled data set was 92.4% for LDA and 85.6% for LR. Based on individual cow LDA, the time spent foraging averaged between 443 and 475 min · d−1. Our results indicated that LCEX can be used to identify the foraging activity of cattle.