Changes in butterfly populations are routinely monitored using transect counts, for example from the UK Butterfly Monitoring Scheme. However, abundance trends are typically only calculated at national and country level in the UK. A new method is presented that estimates species’ trends for smaller regions or datasets, where there may be limited transects. National approaches rely on larger numbers of transects and estimate flight periods either at site level, at the cost of excluding data, or assume a fixed flight period across transects to maximise data usage. The new approach uses butterfly records from all available sources to estimate a parameterised curve representing the flight period and create a so-called Dummy site. Counts from the Dummy site are included with true transect counts in a generalised additive model to estimate annual flight periods as fixed across sites, from which counts and abundance indices are estimated. Inclusion of the Dummy site produces a better overall fit, with greater influence for species with limited transects. Regional indices were often comparable with those produced from a national analysis, but with more realistic indices for some species. Trends were usually similar in magnitude and sign, but for certain species the new approach estimated more robust trends, benefiting from the inclusion of more data through estimating a common regional flight period. The approach is demonstrated for butterflies in Surrey (UK) but has wider relevance, for example to newly-established or small-scale monitoring schemes which may exploit alternative data sources to inform species’ flight period estimation.

中文翻译：

---

一种计算小区域蝴蝶丰度趋势的新方法

蝴蝶种群的变化通常使用横断面计数进行监测，例如来自英国蝴蝶监测计划。然而，丰度趋势通常仅在英国的国家和国家层面计算。提出了一种新方法，可以估计较小区域或数据集的物种趋势，其中可能存在有限的横断面。国家方法依赖于更多的样带并在站点级别以排除数据为代价估计飞行周期，或者假设横断面的固定飞行周期以最大限度地利用数据。新方法使用所有可用来源的蝴蝶记录来估计代表飞行周期的参数化曲线，并创建一个所谓的虚拟站点。来自虚拟站点的计数与真实横断面计数一起包含在广义可加模型中，以估计跨站点固定的年度飞行周期，从中估算计数和丰度指数。包含虚拟站点会产生更好的整体拟合，对横断面有限的物种影响更大。区域指数通常与国家分析得出的指数具有可比性，但某些物种的指数更为现实。趋势通常在幅度和符号上相似，但对于某些物种，新方法估计了更强劲的趋势，受益于通过估计共同区域飞行期而包含的更多数据。这种方法在萨里（英国）的蝴蝶中得到了证明，但具有更广泛的相关性，