Dual-Polarisation Weather Radar for Advanced Monitoring of Aerial Biodiversity (BioDAR)

BioDAR项目将通过利用下一代天气雷达的力量，彻底改变我们记录生活在空气中的动物的丰富性和多样性的方式。气象雷达每5分钟扫描一次整个英国，世界各地都在使用类似类型的雷达。这些雷达通常会探测到蜜蜂和其他昆虫，但由于气象学家对动物不感兴趣，它们被当作不必要的“噪音”而丢弃。这种“噪音”是昆虫多样性和丰富性信息的真正宝库，但我们需要的是一种将雷达所看到的与我们希望监测的昆虫联系起来的方法。BioDAR项目汇集了领先的生态学家和雷达科学家，合作开展一项工作计划，该计划将产生、测试和传播计算机算法，将雷达噪声转化为高质量的生物数据，有可能使我们监测环境的方式发生重大变化。在项目的第一阶段，我们将使用计算机扫描技术，可以对人类头发宽度的十分之一的物体进行成像，以产生60种不同形状的昆虫的高分辨率3D模型。利用物理学中的软件技术，我们可以模拟当这些动物中的每一个通过雷达波束时雷达可能看到的东西。这些模拟的结果将被用于生成算法，该算法可以根据雷达数据的形状将结果分类为不同种类的昆虫，并量化通过波束的昆虫的多样性和数量。在该项目的第二阶段，我们将通过将我们的雷达预测与三个不同的数据集进行比较来测试分类算法。首先，我们将研究三个现有的数据集，这些数据集使用（i）称为“垂直雷达”的特殊雷达来扫描天空的小区域，（ii）每天捕获昆虫的18个吸力陷阱网络，以及（iii）捕获夜间飞蛾的83个灯光陷阱网络。接下来，我们将在12米至1公里的高度范围内使用网进行昆虫采样，并将网连接到气球上。最后，我们将尝试在雷达波束中产生我们自己的昆虫组合，使用实验室饲养的青头苍蝇来饱和雷达周围不同位置的空气。这三个测试将帮助我们了解我们的算法在该领域的表现。在项目的第三阶段，我们将联合收割机结合我们在第一和第二阶段中学习到的分类算法的经验教训，制作一张全国空中昆虫生物多样性和丰度的地图。这张地图将被用来调查保护中的一个紧迫问题：人类对景观的改造对昆虫的影响。我们将从三个方面来研究这个问题，即光污染、城市化和农业环境计划（旨在帮助自然保护农田）的影响。我们预计，在夜间光污染严重、城市化程度较高以及缺乏农业环境规划的地区，昆虫的生物多样性和丰度较低。该项目的最后部分将利用我们所学到的一切（第一阶段的分类算法、第二阶段的验证研究以及第三阶段的国家测绘数据），并将其提供给所有研究人员和公众。我们将使我们所有的数据和分析透明化，以便任何研究人员都可以复制这项工作，我们希望这将使其他国家能够利用我们的发现，将BioDAR方法应用于他们自己的天气雷达网络。这些数据还将被转化为一个在线门户网站，公众可以通过类似于天气预报的界面查看昆虫的生物多样性和丰度。最终的数据集将引起一系列最终用户的极大兴趣，包括地方和国家政府、农民和保护团体。

The BioDAR Project will revolutionise the way in which we record the abundance and diversity of animals that live in the air, by harnessing the power of next generation weather radar. Weather radar scan the entirety of the UK every 5 minutes, and similar types of radar are used around the world for the same purpose. These radar routinely detect bees and other insects, but since animals are not of interest to meteorologists, they are discarded as unwanted "noise". That "noise" is a veritable treasure trove of information on insect diversity and abundance, but what is required is a way to link what a radar sees to the insects that we wish to monitor. The BioDAR Project brings together leading ecologists and radar scientists to collaborate on a programme of work that will produce, test, and disseminate computer algorithms to turn radar noise into high quality biological data with the potential to produce a step change in the way in which we monitor the environment.In the first phase of the project, we will use computer scanning techniques that can image objects 1/10th the width of a human hair to produce high resolution 3D models of a range of 60 different insects of different shapes. Using software techniques from physics, we can simulate what the radar might see when each of those animals passes through the radar beam. The results of those simulations will be used to produce algorithms that can classify results from the radar data into different kinds of insects based on their shape, as well as quantifying the diversity and number of insects passing through the beam.In the second phase of the project, we will then test the classification algorithms by comparing our radar predictions against three different datasets. First, we will look at three existing datasets that have used (i) special radar called "vertical looking radar" to scan small areas of sky, (ii) a network of 18 suction traps that capture insects every day, and (iii) a network of 83 light traps that catch nocturnal moths. Next, we will conduct our own insect sampling using nets at a range of heights from 12m to 1km attached to balloons. Finally, we will attempt to produce our own insect assemblage in the radar beam using lab-reared bluebottle flies to saturate the air in different locations around the radar. These three tests will help us to understand how our algorithms perform in the field.In the third phase of the project, we will combine the lessons learned about our classification algorithms in the first and second phases to produce a national map of aerial insect biodiversity and abundance. This map will be used to investigate a pressing issue in conservation: the effect of human modification of the landscape on insects. We will examine this issue in three ways, by looking at the impacts of light pollution, urbanisation, and agri-environment schemes (which are designed to help nature on farmland). We would expect lower insect biodiversity and abundance near areas with high nocturnal light pollution, higher intensity of urbanisation, and in the absence of agri-environment schemes.The final part of the project will take everything that we have learned (the classification algorithms from phase 1, the validation studies in phase 2, and the national mapping data from phase 3, and make them available to all researchers and the general public. We will make all of our data and analysis transparent so that any researcher can replicate the work, which we hope will enable other countries to make use of our findings to apply the BioDAR approach to their own weather radar networks. The data will also be turned into an online portal which can be accessed by the general public to see insect biodiversity and abundance in an interface similar to a weather forecast. The final datasets will be of great interest to a range of end users, including local and national governments, farmers, and conservation groups.