基于地统计分析西印度洋黄鳍金枪鱼围网渔获量的空间异质性

西印度洋公海海域是世界上围网黄鳍金枪鱼渔业的主要作业海域之一,根据印度洋金枪鱼委员会1999—2004年的1°×1°的各月黄鳍金枪鱼围网渔获量统计数据,采用地统计方法探索该海域黄鳍金枪鱼渔获量的空间异质性特征及其相关生态动力过程。进行了如下分析:(1)利用GIS制图观察渔获量的时空分布特征,发现其空间格局的变异受到的季节变化和年际变化共同影响,且前者明显强于后者。(2)采用地统计方法计算各月渔获量的空间异质性参数,并按照年际和季节情况分别进行了统计分析,发现渔获量的地统计参数值和变异函数模型有明显的季节和年际差异;渔获量的空间相关距离(变程)平均在1000nm左右,冬季要小于夏季;渔获量的空间变异函数模型主要为相关距离较大且空间依赖性较弱的指数模型;渔获量的空间结构方差比例(平均为65.82%)远大于随机性方差比例(平均为34.18%);渔获量在1°×1°尺度下具有明显的空间自相关性。(3)对地统计参数值和渔获量的相关关系研究,并探讨季节变化下渔获量的空间异质性特征与相关生态动力过程关系,发现各月渔获量随着空间总变异(基台值)增加而增加,两者存在强相关性;各月渔获量和南北和西北-东南向分维数值有一定相关性,意味着海洋动力过程在南北和西北—东南向过程越强,渔获量越低。西印度洋黄鳍金枪鱼围网渔获量的空间变异原因在于季风气候和ENSO循环过程引起的海洋流场、营养盐和温跃层等变化外在因素,以及围网捕捞方式和鱼类的行为方式的内在因素共同导致的。

Geostatistical analysis of spatial heterogeneity of yellowfin tuna (Thunnus albacares) purse seine catch in the western Indian Ocean

Our objective was to assess which environmental factors explain the variation in the distribution and abundance of fishery stocks. We used a novel approach, applying geostatistical tools to explain the spatial heterogeneity of purse seine yellowfin tuna (Thunnus albacares) catch in the western Indian Ocean. Geostatistical tools enable researchers to develop a more thorough understanding of the mechanisms controlling the spatial variation in fishery stocks. In addition, this tool is able to deal with complex correlations between spatial patterns and fishery harvest. The western Indian Ocean is one of the most productive purse seine yellowfin tuna (Thunnus albacares) fisheries in the world. We evaluated the variation in the spatial distribution of catch using geostatistical methods. In addition, we discuss the dynamic ecological processes influencing the spatial heterogeneity of catch. We used records of purse seine harvest of yellowfin tuna (Thunnus albacares) collected by the Indian Ocean Tuna Commission (IOTC). The data were summarized by month for 1°×1° areas between 1999-2004. We found that the spatial distribution of catch varied both between seasons and inter-annuals, but was largest between seasons. We obtained the semivariograms parameters and best-fitting semivariogram models from seine yellowfin tuna catch using geostatistical methods. We observed significant seasonal and inter-annual differences in the semivariogram parameters and the semivariogram models of spatial distribution of catch. The average spatial correlation distance (the ranges in geostatistical) was 1000 nautical miles and the values were smaller in winter than in summer. The best-fitting semivariogram models were primarily exponential and had a longer spatial corresponding distance and lower spatial dependence than other models. The spatial structural variance (mean value was 65.82% of total variance) was significantly higher than the random variance (mean value was 34.18% of total variance). We found that the spatial structure of catch had high spatial autocorrelation at 1°×1° areas scale. We investigated the relationship between the semivariogram parameter values and the catch of purse seine yellowfin tuna and attempted to explain the ecological dynamic processes explaining the spatial heterogeneity in catch. We found a strong, positive linear correlation between monthly catch and the sum of spatial variances (sill values), with a correlation coefficient of 0.930 (P<0.001). The monthly catch was also correlated to the south-northward (correlation coefficient=0.5055, P<0.1) and northwest-southeastward fractal dimension values, suggesting that catch was positively correlated with marine dynamic processes oriented in these two directions. Thus, catch decreased when the environmental process components intensified in these two directions. In summary, a number of external factors contribute to the spatial variation in yellowfin tuna catch in the western Indian Ocean, including marine currents, nutrients, and the thickness of the thermocline, which are influenced by the monsoonal climate and ENSO episodes. In addition, we identified internal factors such as the purse seine fishing methods and fish behavior that also affected the spatial distribution of harvest. Our results suggest that several environmental factors can be used to predict changes in the catch of purse seine yellowfin tuna in the western Indian Ocean. These included seasonal indices, which had a significant influence in the spatial distribution model; ENSO episodes; and the optimal semivariogram model. The selection of environmental factors for the yellowfin tuna stock assessment model should include consideration of vector meridian variables that influence marine environmental processes, such as ocean currents or wind fields.