Cloud is a key modulator of the energy budget in earth-atmosphere system. The macro- and micro-physical parameters of cloud have an important influence on precipitation intensity and distributions in arid region. In order to clarify the macroscopic characteristics of cloud in the hinterland of Taklimakan Desert (TD), the structures of cloud layers were continuously detected at Tazhong station from July 23 to September 30 in 2019 by using the CHM15K ceilometer. The variation characteristics of cloud base height, cloud layer thickness, and total cloud coverage were analyzed over Tazhong station in August and September 2019. The results show that the ceilometer could clearly observe the distributions of cloud layer and rainfall processes over TD. The average cloud base height at Tazhong station was about 4.6 km in August and September 2019.The high and middle level clouds predominated at Tazhong station. And the sum of proportions of high and middle level clouds exceeded 90% in August and September, and the proportion of single-layer cloud was significantly higher than that of multi-layer cloud. The cloud layer thickness at Tazhong station was relatively small, the average thicknesses of the first cloud layer were 402 m and 532 m in August and September, respectively, corresponding cloud layer thickness less than 500 m accounted for 64.2% and 58.8%. This implied that the cloud layers were geometrically thin at Tazhong station in August and September 2019. The results of total cloud coverage indicated that the proportion of all-sky cloud day in August was highest (32.6%), whereas the proportion of cloudless sky in September was highest (40.8%), which suggested that clear-sky days were dominant in September. The results of this study will help to understand the spatial and temporal distributions and evolution features of cloud macroscopic characteristics over the hinterland of TD, which provide invaluable datasets for validation of numerical model simulations and satellite remote sensing products.

The automated soil CO₂/gas flux system (model LI-8100A) was used to continuously observe the soil respiration rate, soil temperature and water content in a semi-arid grassland of Loess Plateau during the growing season (from May to October). The temporal variations of soil respiration were comprehensively analyzed, and the influences of environmental hydrothermal factors on soil respiration were studied. The results are as follows: (1) The diurnal dynamic variations of soil respiration rate were significantly different under diverse weather conditions, and the daily mean value (2.90 µmol·m^-2·s^-1) and variation range (1.73-4.92 µmol·m^-2·s^-1) on sunny days were obviously greater than those on cloudy and overcast days. The average diurnal variations of soil respiration rate in different months all showed a significant unimodal pattern, with the maximum (2.20-4.40 µmol·m^-2·s^-1) appearing at 12:00 BST or 13:00 BST and the minimum (0.71-1.70 µmol·m^-2·s^-1) appearing at 05:00 BST or 06:00 BST. The daily mean values were close to the observed values at 10:00 BST or 19:00 BST. (2) The soil respiration rate in both daytime and nighttime were low in May and June, and increased gradually from June, then reached the peak value (in daytime 3.31±0.98 µmol·m^-2·s^-1, in nighttime 1.80±0.39 µmol·m^-2·s^-1) in August, afterwards gradually decreased. Furthermore, the lowest value occurred in October (in daytime 1.55±0.55 µmol·m^-2·s^-1, in nighttime 0.81±0.12 µmol·m^-2·s^-1), and soil respiration rates in daytime were always higher than those in nighttime. The nocturnal soil respiration contributed 27.2%-32.4% to the total respiration during the whole growing season. Therefore, the effects of nocturnal soil respiration should be considered in the current carbon cycle models of grassland ecosystem. (3) Soil temperature was the main environmental factor affecting soil respiration rate, but the univariate model of soil temperature was not enough to fully explain the diurnal dynamic changes of soil respiration. The bivariate nonlinear model combined with soil temperature and water content could be better fitted the soil respiration rate and accounted for 74.0% of the variation. (4) The temperature sensitivity indices (Q₁₀) of the entire day, daytime, and nighttime throughout the growing season in 2020 varied from 1.38 to 2.14, from 1.22 to 1.96, and from 0.85 to 1.64, respectively, with the corresponding mean values of 1.58±0.23, 1.41±0.19 and 1.20±0.16. This suggested that replacing the daily mean values of Q₁₀ with daytime averages would result in an underestimation of about 10.8%.