Linking fish growth and climate across modern space and through evolutionary time : otolith chronologies of the Australian freshwater fish, golden perch (Macquaria, ambigua, Percichthyidae)

Abstract

Australian freshwater systems have highly variable flow regimes influenced by El Nino dry- and La Nina wet- phases of the Southern Oscillation climate phenomenon (SOI). A system characterised by strong climate forcing offers opportunities to disentangle the links between fish growth and climate in order to determine how fish populations could respond to changes in flow arising from climate change or anthropogenic activities (water extraction, river regulation, etc). Life histories (an organism's schedules of growth, maturity and longevity) form the very essence of population dynamics and thus are likely to comprise a sensitive suite of characteristics for reflecting differences in the environment and climates that populations have experienced. An ideal way to investigate life-history variation is through the use of otoliths to extract age information and reconstruct growth chronologies. This thesis is novel in that otolith growth chronologies extracted from an extant fish species were firstly validated as environmental recorders of climate over a range of spatial scales (from single reservoir to river system), and then modem life-histories were compared with Pleistocene remains of the same species over truly evolutionary timescales. This PhD studied the otoliths of the Australian freshwater fish golden perch, Macquaria ambigua (Percichthyidae). Golden perch are common and widespread across fifteen degrees of latitude and occur in four major drainage basins across southeastern and central Australia. Given their broad distribution, extant populations currently experience a wide range of climatic regimes and large differences in age and growth were expected. This study firstly established that golden perch otoliths were suitable for reconstructing growth chronologies by validating the three central assumptions that otoliths could be aged accurately, measured precisely and that otolith growth sensitively reflected relative somatic growth. During this validation, a Biochronology macro was developed to improve the precision of otolith annual-growth measurements by ensuring their orthogonality to the direction of otolith growth even with changing annual opaque zone orientations across the otolith section. This Biochronology macro was then applied to golden perch collected from Googong reservoir, New South Wales, which is at the upper altitudinal and lower thermal limit of the biogeographical range for golden perch. Otoliths proved to be excellent archives of the local environmental conditions experienced over 1982 to 1999 as synchronous interannual growth fluctuations were recorded in the somatic and otolith increment width records for fish of all ages. 73 .4% of the inter-annual variation in the population growth chronology was explained by three environmental parameters: fluctuations in Googong reservoir's water level, minimum dissolved oxygen saturations, and the length of the growing season (number of degree days exceeding 20°C). Low water levels, low oxygen saturations and short growing seasons all combined to result in poor fish growth. These linkages between fish growth and local environmental conditions from a single reservoir were then scaled up to determine if there were synchronous fluctuations between widely separated populations in years of strong climate forcing. 1,240 golden perch otoliths were measured from seven populations that were isolated from one another by barriers to dispersal (dams and weirs). Five populations were clustered around the Australian Capital Territory and two populations were located hundreds of river kilometres downstream in New South Wales. The twenty-nine year growth chronology from 1972 to 2000 was highly correlated with the average annual values of the Southern Oscillation Index (SOI) over 1972-1992 but not over the final eight years 1993-2000 (correlations of 0.59 and -0.62 respectively). This highlights that rather than there being a predictable relationship between fish growth and the SOI through time, populations can drift in and out of synchrony dependent upon the relative combination, strength and autocorrelation of sub-components of high frequency climate disturbances. Growth chronologies were then constructed at the opposite climate extreme to the southern Murray-Darling Basin (MDB), focusing on golden perch inhabiting the intermittent rivers of the Lake Eyre Basin (LEB) that run through the deserts of Central Australia. However, similar to the results for golden perch growth in the cooler and less arid MDB, otolith chronologies in the LEB were highly synchronous among individuals and were significantly correlated with annual discharge. Annual discharge explained 74% and 18% of the inter-annual growth variation for fish from the Diamantina River and Cooper Creek respectively. Finally, the age and growth of golden perch at the two extremes of their extant range (MDB and LEB) were compared with Pleistocene remains of the same species 20- 17,000 years ago (calendar years). This was in order to estimate life-history divergence across modem populations as well as investigate the extent of change over thousands of years. The Pleistocene otoliths were from the dry Willandra Lakes in the MDB, dated back to an oscillatory lake phase shortly after the last glacial maximum (6-9°C cooler than modem). Although there have been no changes to external otolith dimensions, substantial age differences were described for fossil and modem samples. In particular, older age classes were far more common for Pleistocene golden perch (up to 43 years) when compared to their modem counterparts (up to 26 years). Otolith increment widths at age were also narrower, interpreted as representing slower somatic growth during the glacial climate phase. The slower growing and greater longevity Pleistocene golden perch could well have reached larger asymptotic sizes than historical records (23kg and 76cm) given the well known demographic trade-offs among these traits. For comparisons among modem populations, age distributions were similar for LEB and MDB populations, even though the lower mortality MDB environments were expected to have a greater frequency of older age classes. It is suggested that a lack of older age classes in the MDB may represent a legacy of commercial over-exploitation during the 19th and 20th Centuries as well as a disruption of ecological processes through river regulation. Comparisons of otolith growth patterns revealed that the LEB population had wider increment widths-at-age when compared to the MDB, most likely as a function of longer growing seasons, warmer temperatures and faster growth. Spatial divergence of modem life-histories were identified, but the magnitude of differences across modem climate extremes were nowhere near as substantial as the growth and longevity changes over the last 20,000 years of climate change. In conclusion, otoliths offer considerable potential as growth archives to retrieve the environmental histories of both individuals and populations over a wide range of spatial and temporal scales. Show more