Book Description

Millions of dollars are spent controlling weeds in natural ecosystems all over the world, but does the control necessarily result in conservation gain, such as an improvement in native plant recruitment or an increase in the native invertebrate population? The aim of this review was to locate and summarise studies that investigate this question, and attempt to identify any common cause behind the success or failure of weed control programmes. We report on 51 weed removal experiments from 48 studies (15 from New Zealand), assessing the effects of weed control on some aspect of native biodiversity. Thirty-eight experiments compared the effects of one or more treatment methods with untreated areas where the weed population remained intact. Thirteen of the experiments measured the effects of one or more treatment methods, but did not compare results with untreated areas. Almost two-thirds of the experiments had at least one treatment that resulted in positive conservation gain. However, only nine of the experiments reported positive conservation gains for all treatments. Thirty-five percent of the experiments included at least one treatment that failed to achieve any positive conservation gain. The reasons for failure were not always apparent, but regeneration of the target weed, invasion by other weeds and native seed limitation appear to be common factors limiting success. We also looked for studies that investigated whether different levels of weed control resulted in different outcomes for native biodiversity, but found very few. Many studies report a negative linear relationship between weed abundance and native species richness and/or cover. However, statistical methods used were often inadequate, and a non-linear threshold relationship was apparent from some studies, so it is uncertain how much weed populations need to be reduced by in order to achieve conservation gain. In summary, the studies we reviewed were highly variable in terms of the species and ecosystems studied, experimental designs and statistical methods used, and outcomes achieved. Accordingly, we were unable to identify universal factors that determine whether weed control programmes achieve conservation gain.