Root turnover may have substantial implications for nutrient and carbon cycling and for plant breeding. Turnover was calculated for a previously published dataset for five types of perennial ryegrass (Lolium perenne L.) grown in the Waikato, New Zealand. Net root populations were measured with mini-rhizotrons at 2-week intervals for 2 years under well-watered conditions. Measurements were also made for 1 year on well-watered plots (W) or plots not watered (NW) during the summer. We expect the measured root counts to relate more closely to root length than to biomass.

When ample water was supplied, annual gross root production was ∼8 times the average net population, with few differences between the ryegrass types. The general response to withholding water was increased root growth, followed by increased death, and after 3 months there was no substantial net difference between the W and NW treatments. After watering was resumed there was more root growth and death in the NW treatments in the late autumn and winter. The results highlight the importance of making long time-series measurements—differences between watering treatments and depths were sometimes quite different early and late in the experiments. Two ryegrass types had a conspicuous response to water stress, by increasing new root growth at depth, although death rates increased soon after. More research is needed to check for such differences between breeding lines, which might be exploited to improve pasture production in areas prone to drought.