Abstract:

The city belt along the Yellow River in Ningxia is located at the northeast of Ningxia Province,China.It is both a key area for geographical research in arid areas and the core of economic development in Ningxia.As an important climatic phenomenon related to the process of urbanization,the heat island effect has brought much inconvenience to human life,for which reason it has become a major topic in current urbanization research.Most previous studies have been conducted on wet climate regions,and there are relatively few studies on arid and semi-arid regions.Moreover,the time series used in previous studies has typically been short and their resolution low,hence the analysis of urban-heat-island trends and the reasons for their occurrence are not yet comprehensively understood.Based on MODIS surface temperature products,this paper uses the Mann-Kendall non-parametric test and Sen’s slope estimation method to analyze the time trend of surface temperature during the past 20 years along city belt of the Yellow River.Combining MODIS surface temperature data with MODIS land cover type products,and introducing the urban-heat-island ratio index,daily and seasonal changes in the urban-heat-island effect along the city belt and their time series trends over the past 20 years were analyzed.The results show that the surface temperatures of most areas along the city belt in Ningxia have shown an upward trend,which is stronger during nighttime than daytime.An increase of surface coverage is an important reason for this trend.Compared with humid climates,the heat island effect along the city belt shows different daily and seasonal trends.At nighttime,the heat island effect is stronger than during the daytime and in over course of one year,the heat island effect is stronger in winter and spring at daytime but stronger in summer at nighttime.The heat island effect in the study areas shows a decreasing trend,whereas the nighttime heat island effect show an increasing trend,but the trend was not found to be significant.Finally,land and vegetation cover types were analyzed.To some extent,our results indicate that urbanization and surface coverage are equally important contributions to heat islands along the study area.Surface coverage increases during the daytime can weaken surface temperatures increases,while promoting such increases at night.Increased surface coverage can thus alleviate increases in surface temperature during the daytime.Urbanization does not necessarily promote the heat island effect during daytime,but sometimes mitigates its effect on low vegetation coverage areas,such as bare land.Nonetheless,the impact of urbanization on the heat island effect is more obvious at night.This study provides a direction for the management of arid areas and a reference for studies of temporal and spatial changes in surface temperature due to the heat island effect in other arid regions.It also gives guidance for studies concerning the influence of such phenomena in other arid regions．

Key words: urban-heat-island ratio index, Mann-Kendall nonparametric test, Sen’s slope estimation, surface cover type, surface vegetation coverage