Comparison of damage assessment methods of coconut rhinoceros beetle (Oryctes rhinoceros): A thesis submitted in partial fulfilment of the requirements for the Degree of Master of Science at Lincoln University

Oryctes rhinoceros (L.) (Coleoptera: Scarabaeidae), most commonly known as the coconut rhinoceros beetle or CRB, is an invasive pest that affects palms, especially coconut (Cocos nucifera) and oil palm (Elaeis guineensis) and their effects can be detrimental. For the past 40 years, CRB population has been controlled by the biological control Oryctes nudivirus (OrNV), however, a recent incursion causing adverse impacts in the Pacific has prompted researchers to re-evaluate the existing management approaches, so effective population control can be achieved once again. A vital component of these management process that is often overlooked, is the need for a robust damage assessment and monitoring tool. Various damage assessment methods have been developed and the use of damage scales has been the most common among them, however, their accuracy and potential to monitor changes over time have not been explored. This study aimed to achieve two main objectives: 1) To compare and identify the capability of 3 damage scales, binary, 3-point and 5-point, in assessing damage severity levels and monitoring changes that occur over time and 2) To determine whether assessors affected the accuracy of the 3 damage scales. Photographs from a historical collection of CRB studies done over the years in the Solomon Islands, Papua New Guinea and Fiji were used to test the capabilities of the scales. These countries were selected due to their differing CRB status. The influence of assessors was tested with data that was collated from an online survey with respondents including those with no knowledge of CRB and experts in the field. The results indicated that damage intensity levels can be measured through the 5-point scale, but the accuracy within the scale is low compared to the binary and 3-point scales. The study also found that detecting changes over time is feasible however, it was difficult to determine the minimum detectable effect size. A brief simulation study found that the effect size in a sample of 100 undamaged palms is at 6% in the binary and 3-point scale. It was recommended that for effective monitoring of changes a sample population of 100 or more palms is required. Given that change was detectable after a minimum of 2 years, it is practical that monitoring periods are scheduled biannually. The online survey revealed that experience increases the accuracy of the damage scales among assessors, nevertheless, training still remains a vital requirement in the use of the scales if accurate and reliable results are expected.