The implications of increasing variability of fish landings

Abstract

Ghoti papers Ghoti aims to serve as a forum for stimulating and pertinent ideas. Ghoti publishes succinct commentary and opinion that addresses important areas in fish and fisheries science. Ghoti contributions will be innovative and have a perspective that may lead to fresh and productive insight of concepts, issues and research agendas. All Ghoti contributions will be selected by the editors and peer reviewed. Etymology of Ghoti George Bernard Shaw (1856‐1950), polymath, playwright, Nobel prize winner, and the most prolific letter writer in history, was an advocate of English spelling reform. He was reportedly fond of pointing out its absurdities by proving that ‘fish’ could be spelt ‘ghoti’. That is: ‘gh’ as in ‘rough’, ‘o’ as in ‘women’ and ‘ti’ as in palatial.AbstractIn this paper, we measured the variance of fisheries landings using 344 time series from the Food and Agriculture Organization (FAO) of the United Nations: the series chosen had neither blank records nor zeroes over the period 1950–1996. We compared the results with 431 terrestrial populations from the Global Population Dynamics Database. The logarithm of landings (for fish) and the logarithm of population abundance (for terrestrial organisms) were employed. For both original and detrended time series, the Hurst exponent (a time series autocorrelation index that detects non‐random changes) was used to measure the rate of increase of variance as a function of time span. For most FAO time series, variability increased with series length. A large proportion of fish landings series had very large Hurst exponents, suggesting a greater contribution to the total variance by longer‐period oscillations for marine data relative to terrestrial ones but this difference between the marine and terrestrial data virtually disappeared when the series were detrended. In fish landings and terrestrial series there was no indication of variance stabilizing at a finite value. This growth in variance may extend to longer time scales, as indicated by analysis of two of the longest time series for marine fish. Our findings have major ecological and management implications as continuously increasing variability implies that there can be no long‐term equilibrium yield. This is incompatible with many models currently used for fisheries management that assume bounded equilibrium yields. Models or management strategies looking to the long‐term must account for this phenomenon.