Non-detection errors in a survey of persistent, highly-detectable vegetation species

Area of Science:

• Ecology
• Botany
• Conservation Biology

Background:

• Conventional vegetation surveys often fail to detect rare, small, or annual species.
• Detectability of common, persistent vegetation species is assumed high but lacks empirical support.
• Non-detection errors are a well-established issue in fauna surveys but understudied in vegetation science.

Purpose of the Study:

• To quantify false-negative errors in detecting perennial vegetation species using systematic surveys in arid South Australia.
• To assess the impact of observer skill on detection probabilities.
• To predict non-detection error rates by comparing different survey methodologies.

Main Methods:

• Focused on seven easily detectable perennial species in arid South Australia.
• Controlled for observer skill during systematic vegetation surveys.
• Compared methodologies to estimate non-detection error rates for a second survey.

Main Results:

• All but one highly detectable perennial species exhibited imperfect detection.
• Detection probabilities ranged significantly from 0.22 to 0.83.
• Non-detection error rates were predicted for a secondary survey based on methodological comparisons.

Conclusions:

• Systematic vegetation surveys may yield flawed inferences if non-detection errors are not identified and addressed.
• The study underscores the critical need to acknowledge and manage non-detection issues in vegetation science.
• Adopting fauna survey techniques for non-detection is recommended for vegetation surveys.