Landscape-land surface temperature relation in an arid hot city

The research hypothesis is that detailed landscape metrics when correlated with diurnal and nocturnal land surface temperature (LST) and analysed with linear regression can help identify heat reduction policies for hot arid cities.
The data show that in Africa, cities in hot arid zones are poorly investigated, that the urban-rural gradient of LST is controversial, landscape features most related to urban heat are highly variable, and that recommended policies are mainly oriented towards increasing vegetation but remain generic. In Niamey (Niger) road width, tree and shrub cover are the most correlated landscape metrics with diurnal LST. Road width is highly correlated with nocturnal LST. The other metrics are not statistically significant. The road width-nocturnal LST relationship is linear.
The dataset contains six tables and three figures. The first three tables cover (1) published studies of land surface temperature (LST) in African cities with more than 100,000 inhabitants per climate zone, (2) the main Landscape-LST dynamics observed, and (3) the policies recommended by published literature to reduce LST. Full references to the studies considered are provided. The last three tables contain data on the Landscape-LST relationship in the city of Niamey, Niger. The remaining three tables show (4) the eight landscape metrics and the average daytime and nighttime LST observed in 29 neighbourhood samples in May 2020, (5) the correlation matrix, and (6) the stepwise linear regression results are provided. The three figures represent (1) the location of the neighbourhood samples considered and (2-3) the normal probability plot of residuals.
The eight landscape metrics are collected by visual photointerpretation of very high-resolution Google Earth Pro images from May 2020. Daytime and nighttime LST are extracted from ECOSTRESS data from May 2 and 6, 2020.
This dataset can be used as a baseline against which to ascertain changes at later dates or to compare landscape features and LST with those of other cities in the same arid hot climate (Bsh, according to the Köppen-Geiger classification).

本研究的研究假设为：将精细化景观指数与昼夜地表温度（Land Surface Temperature，LST）进行关联，并通过线性回归分析，可助力炎热干旱城市的降温政策制定。

数据表明，针对非洲炎热干旱区城市的相关研究仍较为匮乏；LST的城乡梯度存在争议；与城市热效应关联度最高的景观特征差异显著；现有推荐政策多以增加植被覆盖为核心，但仍缺乏针对性。

在尼日尔尼亚美市，道路宽度、乔木与灌木覆盖度是与日间LST关联度最高的景观指数；道路宽度与夜间LST的关联度极高，其余景观指数则无统计学显著性；道路宽度与夜间LST之间呈线性相关关系。

本数据集包含6张表格与3幅插图。前3张表格分别涵盖：(1) 按气候区划分的、人口超10万的非洲城市LST相关已发表研究；(2) 已观测到的主要景观-LST动态特征；(3) 已发表文献中提出的LST降温政策。数据集同时提供了所纳入研究的完整参考文献列表。

后3张表格聚焦尼日尔尼亚美市的景观-LST关联关系，具体包括：(4) 2020年5月29个街区采样点的8项景观指数与平均日间、夜间LST观测数据；(5) 相关系数矩阵；(6) 逐步线性回归分析结果。

3幅插图分别为：(1) 本次研究所纳入的街区采样点分布位置；(2-3) 残差的正态概率分布图。

8项景观指数通过对2020年5月的超高分辨率Google Earth Pro影像进行目视解译获取；日间与夜间LST数据则提取自2020年5月2日及6日的ECOSTRESS遥感数据集。

本数据集可作为基准基线，用于后续监测区域变化，或是与其他同属炎热干旱气候（根据柯本-盖格（Köppen-Geiger）气候分类为Bsh型）的城市的景观特征与LST数据进行对比分析。