The catchment land-use composition of 249 fish sampling sites in Austrian running waters revealed effects on the biological integrity. Beyond correlative analysis, we investigated (1) which land-use category had the strongest effect on fish, (2) whether metrics of functional fish guilds reacted differently, (3) whether there were cumulative effects of land-use categories, and (4) whether effects varied in strength across river types. We fed 5 land-use categories into regression trees to predict the European Fish Index or fish metric of intolerant species (mainly Salmo trutta fario). Agriculture and urbanisation were the best predictors and indicated significant effects at levels of >23.3 and >2%, respectively. Model performance was R2 = 0.15 with the Fish Index and R2 = 0.46 with intolerant species. The tree structure showed a cumulative effect from agriculture and urbanisation. For the intolerant species metric, a combination of high percentages for agriculture and urbanisation was related to moderate status, whereas <7.3% agriculture were related to good status, although urbanisation was higher than 1.8%. Headwater river types showed stronger responses to land use than river types of lower gradient and turned out to be more sensitive to urbanisation than agriculture.

中文翻译：

---

土地利用对奥地利不同类型流水鱼类指标的累积影响

奥地利流域 249 个鱼类采样点的集水区土地利用构成揭示了对生物完整性的影响。除了相关分析，我们还调查了 (1) 哪个土地利用类别对鱼类的影响最大，(2) 功能性鱼类公会的指标是否有不同的反应，(3) 土地利用类别是否存在累积效应，以及 (4) ) 不同河流类型的影响强度是否不同。我们将 5 个土地利用类别输入回归树，以预测不耐受物种（主要是鲑鱼）的欧洲鱼类指数或鱼类指标。农业和城市化是最好的预测指标，分别在 >23.3 和 >2% 的水平上显示出显着影响。模型性能对于鱼类指数为 R2 = 0.15，对于不耐受物种为 R2 = 0.46。树状结构显示出农业和城市化的累积效应。对于不耐受物种指标，农业和城市化的高百分比与中等状态相关，而 <7.3% 的农业与良好状态相关，尽管城市化率高于 1.8%。源头河流类型对土地利用的反应比低坡度的河流类型更强，并且比农业对城市化更敏感。