Spatial analysis of changing terrestrial ecosystems in the Windmill Islands and the sub-Antarctic

Metadata record for data from AAS (ASAC) project 3130.Public High latitude terrestrial ecosystems are experiencing rapid change, which is most likely caused by climate change, human impacts, and invasive species. Up-to-date and accurate spatial data at a range of scales are of crucial importance for mapping changes in these fragile ecosystems. The aim of this study is to undertake spatial analyses on the changing terrestrial ecosystems of the Windmill Islands, Antarctica and sub-Antarctic Macquarie Island. The study aims to better understand the different processes that result in ecosystem change and with new state-of-the-art high-resolution spatial data we hope to contribute to improved management strategies.Project Objectives:IntroductionEnvironmental threats globally can be categorised into four main types: local impact from human activity and habitat loss; impact from alien species and homogenisation of biota; impact from climate change and impact associated with harvesting and resource extraction. All four types of impacts occur to some degree in the Antarctic region (Hull and Bergstrom 2006, Bergstrom and Selkirk 2007). This project examines change associated with these impacts in Australian Antarctic and sub-Antarctic territories. In particular, we seek to isolate signals of impact from regional climate change from those of other human-induced change within Antarctic and sub-Antarctic terrestrial ecosystems.This project will develop and apply spatial data collection and analysis techniques for detailed baseline mapping and change detection of vegetation communities on the Windmill Islands and Macquarie Island. We will then employ these cutting-edge techniques to quantify, detect, and understand the impact of changes. In detail, the objectives of this project are to:Objective 1: Collate and collect spatial data in order to establish a baseline map of, and detect changes in vegetation communities on the Windmill Islands and Macquarie Island.Objective 2: Create high-resolution digital elevation models (DEM) based on GPS data and airborne laser scanning (LiDAR) of the localities.Objective 3: Explore ecological relationships between vegetation communities and biologically relevant landscape characteristics and human-induced disturbance using terrain analysis of digital elevation models in a Geographical Information System (GIS) in order to better understand the distribution of and changes in vegetation communities. This will include the development of hydrological terrain analyses to examine the impact of changing snow conditions around Casey on vegetation communities.Objective 4: Develop and apply new multi-scale field sampling techniques based on field photogrammetry and GPS observations at different scales (from 20cm to 20m) to measure relative percent cover of plant species and vegetation communities. This objective is of key importance to bridge the range of scale levels from small field quadrats to satellite images that cover large portions of the landscape.Objective 5: Combine detailed plot-scale data and field photographs with terrain information and high-resolution satellite imagery to identify and map changes in both plant communities and plant stress more efficiently.This project will deliver valuable baseline and temporal data on the impact of environmental change in Australian Antarctic and sub-Antarctic territories. It will improve our understanding of Antarctic and sub-Antarctic landscape ecology and species adaptations. It will provide a predictive GIS model that can forecast the effects of human activities in Antarctica and provide new tools for spatial multi-scale geographic analysis. Taken from the 2009-2010 Progress Report:Progress against objectives:Objective 1:Windmill Islands moss bedsIn the first year of this project we found that the spatial scale of the moss beds (tens of m2) makes satellite imagery (even very high resolution imagery of 0.5 m) unsuitable for mapping their extent in sufficient detail. Due to logistical constraints aerial photography is impractical. Recent developments in the use of unmanned aerial vehicles (UAVs) for remote sensing applications provide exciting new opportunities for ultra-high resolution mapping and monitoring of the environment. This year, we developed a new UAV consisting of an electric remote controlled helicopter capable of carrying three different cameras: visible colour, near-infrared, and thermal infrared for cost-effective, efficient, and ultra-high resolution (less than 5 cm pixel size) mapping of terrestrial vegetation in the Windmill Islands, Antarctica. These three sensors allowed us to map different physical characteristics of the moss beds at resolutions of several centimetres.We had a very successful season at Casey. We managed to collect spatial data for four different moss sites: ASPA135, Red Shed, Robinson Ridge, ASPA136. We collected the following datasets:- Very accurate GPS locations for existing moss quadrat sites with a geodetic GPS receiver (cm accuracy).- For ASPA135, Red Shed, and Robinson Ridge we collected very dense GPS transects and used these data to interpolate high resolution digital elevation models (DEMs).- For all sites we collected geotagged photographs of all quadrats in addition to geotagged landscape scale photographs.- For ASPA135, Red Shed, and Robinson Ridge we flew a total of 26 UAV flights collecting visible photography (2 cm pixel size), near-infrared photography, thermal imagery, and video footage for all sites.- For the Robinson Ridge and Red Shed site we collected spectral signatures of the key moss species and other land cover types (water, rock types, lichen, snow, etc.). The handheld spectrometer was rented from Geoscience Australia.- On request of Sandra Potter and Tom Maggs, we collected GPS data and UAV photography for the Casey quarry before and after blasting to determine the extent of the blasting zone and to acquire ultra-high resolution imagery of the quarry for management purposes.Macquarie IslandThis project has strong links with AAS project 3095. Phillippa Bricher (UTAS PhD student) and Jared Abdul-Rahman (UTAS volunteer and Honours student) have collected data for Phillippa's PhD project. Data collection for Phillippa's project consisted of geotagged photographs of vegetation plots with Polecam. Jared concentrated on photographing Azorella die-back. Phillippa's data will be used for vegetation classification of the island using satellite imagery and DEMs.A new WorldView-2 high-resolution satellite image was acquired for the northern half of the island on 26 December 2009. This image will be extremely useful for vegetation classification and change detection.Objective 2As noted in objective 1 (above), we collected dense transects of GPS data for three moss bed sites in the Windmill Islands. We interpolated the GPS height values to obtain three very high resolution DEMs (less than 0.5 m). The AAD's LiDAR instrument was not available at Casey or Macquarie Island this season, however, we requested LiDAR data collection at Davis over known moss sites. The data was collected successfully, but it hasn't been processed yet. With this dataset we are hoping to assess the usefulness of LiDAR for mapping of micro-topography. In the meantime we have continued to develop our UAV (externally funded UTAS project). We have built a larger version that is capable of carrying a mini-LiDAR instrument. We hope to employ this UAV LiDAR at our study sites in the Windmill Islands during the 2010/2011 summer season. This novel system will allow us to capture the microtopography of the moss bed areas and will allows us to more accurately model the hydrological conditions (compared to GPS derived DEMs).Objective 3:We have already modelled several environmental parameters for the high-resolution DEMs of the Windmill Islands (ASPA135, Robinson Ridge, and the Red Shed). The derivatives include a topographic wetness index, average annual solar radiation, and slope gradient. In combination with the UAV photographs and the close-up quadrat photographs we aim to establish a relationship between the condition of the moss and environmental factors.Lucieer is currently on Study Leave at ITC in The Netherlands (March - April 2010) and the University of Calgary, Canada (April - May 2010). At these institutes Lucieer is working on a new texture-based classification technique to map healthy tussock slopes on Macquarie Island (as an indicator of island health). Preliminary highlight that this novel image classification technique is very successful at identifying tussock slopes in high resolution QuickBird imagery.Objective 4:With the Polecam technique on Macquarie Island and with the UAV photographs in the Windmill Islands we have developed two very novel techniques for multi-scale sampling. These photographic sampling techniques will provide invaluable information for the next phase of the project.Objective 5:We aim to further develop our UAV project and use the larger UAV with multiple sensor in future field campaigns. This will allow us to build a multi-temporal dataset of the study areas and detect changes over time. The experiments in this first field season have provided us with important insights for suitable data collection techniques and the collected data are incredibly valuable for addressing the objectives of this project.

AAS（ASAC）项目3130数据元记录

公开

高纬度陆地生态系统正经历快速变化，这很可能由气候变化、人类活动影响及入侵物种所致。多尺度的最新准确空间数据对于绘制这些脆弱生态系统的变化图谱至关重要。本研究旨在对南极洲风车群岛及亚南极麦夸里岛的变化中陆地生态系统开展空间分析，以更好地理解导致生态系统变化的各类过程；借助最新的高分辨率空间数据，我们期望为优化管理策略提供支持。

项目目标：
引言

全球环境威胁可分为四类主要类型：人类活动及栖息地丧失的局部影响、外来物种及生物群同质化的影响、气候变化的影响，以及与资源采集和开采相关的影响。这四类影响在南极地区均有不同程度的存在（Hull and Bergstrom 2006, Bergstrom and Selkirk 2007）。本项目考察澳大利亚南极及亚南极领地内与这些影响相关的变化，尤其旨在从南极及亚南极陆地生态系统的其他人为变化中分离出区域气候变化的影响信号。

本项目将开发并应用空间数据采集与分析技术，用于风车群岛及麦夸里岛植被群落的详细基准制图与变化检测。随后，我们将运用这些前沿技术对变化的影响进行量化、检测与理解。具体而言，项目目标如下：

目标1：整理并收集空间数据，以建立风车群岛及麦夸里岛植被群落的基准图，并检测其变化。

目标2：基于各地点的全球定位系统（GPS）数据及机载激光扫描（LiDAR）数据，创建高分辨率数字高程模型（DEM）。

目标3：利用地理信息系统（GIS）中的数字高程模型地形分析，探索植被群落与生物相关景观特征及人为干扰之间的生态关系，以更好地理解植被群落的分布及变化。这将包括开发水文地形分析方法，以研究Casey站周边积雪条件变化对植被群落的影响。

目标4：基于不同尺度（20厘米至20米）的野外摄影测量及GPS观测，开发并应用新型多尺度野外采样技术，以测量植物物种及植被群落的相对覆盖百分比。该目标对于衔接从小型野外样方到覆盖大片景观的卫星影像之间的尺度范围至关重要。

目标5：将详细的样方尺度数据、野外照片与地形信息及高分辨率卫星影像相结合，更高效地识别并绘制植物群落变化及植物胁迫状况。

本项目将提供澳大利亚南极及亚南极领地内环境变化影响的宝贵基准数据与时间序列数据，提升我们对南极及亚南极景观生态学与物种适应性的理解。项目还将构建一个预测性GIS模型，用于预测南极洲人类活动的影响，并提供空间多尺度地理分析的新工具。

摘自2009-2010年度进展报告：
进展情况：
目标1：
风车群岛苔藓床
项目第一年，我们发现苔藓床的空间尺度（数十平方米）使得卫星影像（即使是0.5米的超高分辨率影像）也无法足够详细地绘制其范围。受后勤限制，航空摄影亦不切实际。无人机（unmanned aerial vehicle, UAV）在遥感应用中的最新发展，为环境的超高分辨率制图与监测提供了令人振奋的新机遇。本年度，我们研发了一款新型无人机——由电动遥控直升机构成，可搭载三种不同相机（可见光彩色相机、近红外相机及热红外相机），用于对南极洲风车群岛的陆地植被进行高性价比、高效且超高分辨率（像素尺寸小于5厘米）的制图。这三种传感器使我们能够以数厘米的分辨率绘制苔藓床的不同物理特征。

我们在Casey站度过了非常成功的考察季，为四个不同的苔藓站点收集了空间数据：ASPA135、Red Shed、Robinson Ridge及ASPA136。收集的数据集如下：
- 使用大地测量GPS接收机（厘米级精度）获取现有苔藓样方站点的高精度GPS位置。
- 针对ASPA135、Red Shed及Robinson Ridge，收集了高密度GPS样带数据，并利用这些数据插值生成高分辨率数字高程模型（DEM）。
- 针对所有站点，除了带地理标记的景观尺度照片外，还收集了所有样方的带地理标记照片。
- 针对ASPA135、Red Shed及Robinson Ridge，共进行了26次无人机飞行，为所有站点收集了可见光照片（像素尺寸2厘米）、近红外照片、热影像及视频 footage。
- 针对Robinson Ridge及Red Shed站点，收集了关键苔藓物种及其他土地覆盖类型（水、岩石类型、地衣、雪等）的光谱特征。手持光谱仪从澳大利亚地质科学局（Geoscience Australia）租赁。
- 应Sandra Potter与Tom Maggs的请求，我们在Casey采石场爆破前后收集了GPS数据及无人机照片，以确定爆破区域的范围，并获取采石场的超高分辨率影像用于管理目的。

麦夸里岛
本项目与AAS项目3095联系紧密。Phillippa Bricher（塔斯马尼亚大学（UTAS）博士生）及Jared Abdul-Rahman（UTAS志愿者及荣誉学生）为Phillippa的博士项目收集了数据。Phillippa项目的数据收集包括使用Polecam拍摄植被样方的带地理标记照片；Jared则专注于拍摄Azorella枯梢病（die-back）的照片。Phillippa的数据将用于利用卫星影像及DEM对该岛进行植被分类。

2009年12月26日，获取了该岛北半部的新WorldView-2高分辨率卫星影像，这将对植被分类及变化检测极为有用。

目标2：
如目标1（上文）所述，我们为风车群岛的三个苔藓床站点收集了高密度GPS样带数据，并对GPS高程值进行插值，得到了三个超高分辨率DEM（分辨率小于0.5米）。本年度考察季，澳大利亚南极局（AAD）的LiDAR仪器未在Casey站或麦夸里岛部署，但我们请求在Davis站已知苔藓站点进行LiDAR数据采集，该数据已成功收集，但尚未处理。我们希望利用该数据集评估LiDAR在微地形制图中的实用性。同时，我们继续推进无人机项目（由UTAS外部资助），已构建了一款更大的无人机，能够搭载迷你LiDAR仪器。我们期望在2010/2011夏季考察季将这款无人机LiDAR应用于研究站点。该新型系统将使我们能够捕捉苔藓床区域的微地形，并能更准确地模拟水文条件（相较于GPS衍生的DEM）。

目标3：
我们已为风车群岛的三个苔藓床站点（ASPA135、Robinson Ridge及Red Shed）的高分辨率DEM模拟了多个环境参数，包括地形湿度指数、年平均太阳辐射及坡度。结合无人机照片及近距离样方照片，我们旨在建立苔藓状况与环境因素之间的关系。

Lucieer目前正在荷兰ITC（3月-4月）及加拿大卡尔加里大学（4月-5月）进行学术休假。在这些机构，Lucieer正致力于开发一种新型基于纹理的分类技术，用于绘制麦夸里岛的健康草丛斜坡（作为岛屿健康的指标）。初步结果表明，这种新型影像分类技术在高分辨率QuickBird影像中识别草丛斜坡方面非常成功。

目标4：
通过麦夸里岛的Polecam技术及风车群岛的无人机照片，我们开发了两种非常新颖的多尺度采样技术。这些摄影采样技术将为项目下一阶段提供宝贵信息。

目标5：
我们旨在进一步推进无人机项目，并在未来的野外考察中使用搭载多传感器的更大无人机。这将使我们能够构建研究区域的多时相数据集，并检测随时间的变化。首个野外考察季的实验为我们提供了关于合适数据采集技术的重要见解，收集的数据对于实现项目目标具有极高价值。