Homogeneity analysis of the streamflow time series is essential to hydrological modeling, water resources management and climate change studies. In this study, five absolute homogeneity tests and one clustering approach were used to determine the homogeneity status of the streamflow time series (over the period 1960-2010) in 14 hydrometric stations of three important basins (i.e., Aras River Basin, Urmia Lake Basin and Sefid-Roud Basin) in northwestern Iran. Results of the Buishand range test, von Neumann ratio test, cumulative deviation test, standard normal homogeneity test and Pettitt test for monthly streamflow time series detected that about 42.26%, 38.09%, 33.33%, 39.28% and 68.45% of the streamflow time series were inhomogeneous at the 0.01 significance level, respectively. Streamflow time series of the stations located in the eastern parts of the study area or within the Urmia Lake Basin were mostly homogeneous. In contrast, streamflow time series in the stations of the Aras River Basin and Sefied-Roud Basin showed inhomogeneity at annual scales. Based onthe overall classification for the monthly and annual streamflow series, we determined that about 45.60%, 11.53% and 42.85% of the time series were categorized into the 'useful', 'doubtful' and 'suspect' classes according to the five absolute homogeneity tests. We also found the homogeneity patterns of the streamflow time series by using the clustering approach. The results suggested the effectiveness of the clustering approach for homogeneity analysis of the streamflow time series in addition to the absolute homogeneity tests. Moreover, results of the absolute homogeneity tests and clustering approach indicated obvious decreasing change points of the streamflow time series in the 1990s over the three basins, which were mostly related to the hydrological droughts.

The Palmer drought severity index (PDSI), standardized precipitation index (SPI), and standardized precipitation evapotranspiration index (SPEI) are used worldwide for drought assessment and monitoring. However, substantial differences exist in the performance for agricultural drought among these indices and among regions. Here, we performed statistical assessments to compare the strengths of different drought indices for agricultural drought in the North China Plain. Small differences were detected in the comparative performances of SPI and SPEI that were smaller at the long-term scale than those at the short-term scale. The correlation between SPI/SPEI and PDSI considerably increased from 1- to 12-month lags, and a slight decreasing trend was exhibited during 12- and 24-month lags, indicating a 12-month scale in the PDSI, whereas the SPI was strongly correlated with the SPEI at 1- to 24-month lags. Interestingly, the correlation between the trend of temperature and the mean absolute error and its correlation coefficient both suggested stronger relationships between SPI and the SPEI in areas of rapid climate warming. In addition, the yield-drought correlations tended to be higher for the SPI and SPEI than that for the PDSI at the station scale, whereas small differences were detected between the SPI and SPEI in the performance on agricultural systems. However, large differences in the influence of drought conditions on the yields of winter wheat and summer maize were evident among various indices during the crop-growing season. Our findings suggested that multi-indices in drought monitoring are needed in order to acquire robust conclusions.

Exploring the surface energy exchange between atmosphere and water bodies is essential to gain a quantitative understanding of regional climate change, especially for the lakes in the desert. In this study, measurements of energy flux and water vapor were performed over a lake in the Badain Jaran Desert, China from March 2012 to March 2013. The studied lake had about a 2-month frozen period (December and January) and a 10-month open-water period (February-November). Latent heat flux (LE) and sensible heat flux (Hs) acquired using the eddy covariance technique were argued by measurements of longwave and shortwave radiation. Both fluxes of longwave and shortwave radiation showed seasonal dynamics and daily fluctuations during the study period. The reflected solar radiation was much higher in winter than in other seasons. LE exhibited diurnal and seasonal variations. On a daily scale, LE was low in the morning and peaked in the afternoon. From spring (April) to winter (January), the diurnal amplitude of LE decreased slowly. LE was the dominant heat flux throughout the year and consumed most of the energy from the lake. Generally speaking, LE was mostly affected by changes in the ambient wind speed, while Hs was primarily affected by the product of water-air temperature difference and wind speed. The diurnal LE and Hs were negatively correlated in the open-water period. The variations in Hs and LE over the lake were differed from those on the nearby land surface. The mean evaporation rate on the lake was about 4.0 mm/d over the entire year, and the cumulative annual evaporation rate was 1445 mm/a. The cumulative annual evaporation was 10 times larger than the cumulative annual precipitation. Furthermore, the average evaporation rates over the frozen period and open-water period were approximately 0.6 and 5.0 mm/d, respectively. These results can be used to analyze the water balance and quantify the source of lake water in the Badain Jaran Desert.

The rapid desertification of grasslands in Inner Mongolia of China poses a significant ecological threaten to northern China. The combined effects of anthropogenic disturbances (e.g., overgrazing) and biophysical processes (e.g., soil erosion) have led to vegetation degradation and the consequent acceleration of regional desertification. Thus, mitigating the accelerated wind erosion, a cause and effect of grassland desertification, is critical for the sustainable management of grasslands. Here, a combination of mobile wind tunnel experiments and wind erosion model was used to explore the effects of different levels of vegetation coverage, soil moisture and wind speed on wind erosion at different positions of a slope inside an enclosed desert steppe in the Xilamuren grassland of Inner Mongolia. The results indicated a significant spatial difference in wind erosion intensities depending on the vegetation coverage, with a strong decreasing trend from the top to the base of the slope. Increasing vegetation coverage resulted in a rapid decrease in wind erosion as explained by a power function correlation. Vegetation coverage was found to be a dominant control on wind erosion by increasing the surface roughness and by lowering the threshold wind velocity for erosion. The critical vegetation coverage required for effectively controlling wind erosion was found to be higher than 60%. Further, the wind erosion rates were negatively correlated with surface soil moisture and the mass flux in aeolian sand transport increased with increasing wind speed. We developed a mathematical model of wind erosion based on the results of an orthogonal array design. The results from the model simulation indicated that the standardized regression coefficients of the main effects of the three factors (vegetation coverage, soil moisture and wind speed) on the mass flux in aeolian sand transport were in the following order: wind speed>vegetation coverage>soil moisture. These three factors had different levels of interactive effects on the mass flux in aeolian sand transport. Our results will improve the understanding of the interactive effects of wind speed, vegetation coverage and soil moisture in controlling wind erosion in desert steppes, and will be helpful for the design of desertification control programs in future.

The measurement and assessment of dust emissions from different landforms are important to understand the atmospheric loading of PM10 (particulate matter≤10 μm aerodynamic diameter) and to assess natural sources of dust; however, the methodology and technique for determining the dust still present significant research challenges. In the past, specialized field observation and field wind tunnel studies have been used to understand the dust emission. A series of wind tunnel tests were carried out to identify natural sources of dust and measure the magnitudes of dust emissions from different landforms. The method used in this study allowed the measurement of the PM10 emission rate using a laboratory based environmental boundary layer wind tunnel. Results indicated that PM10 emissions demonstrated strong temporal variation and were primarily driven by aerodynamic entrainment. Sand dunes, playa, and alluvial fans had the largest dust emission rates (0.8-5.4 mg/(m²?s)) while sandy gravel, Gobi desert and abandoned lands had the lowest emission rates (0.003-0.126 mg/(m²?s)). Dust emissions were heavily dependent on the surface conditions, especially the availability of loose surface dust. High dust emissions were a result of the availability of dust-particle materials for entrainment while low dust emissions were a result of surface crusts and gravel cover. Soil surface property (surface crusts and gravel cover) plays an important role in controlling the availability of dust-sized particles for entrainment. The dust emission rate depended not only on the surface conditions but also on the friction velocity. The emission rate of PM10 varies as a power function of the friction velocity. Although dynamic abrasion processes have a strong influence on the amount of dust entrainment, aerodynamic entrainment may provide an important mechanism for dust emissions. Large volumes of dust entrained by aerodynamic entrainment cannot only occur at low shear velocity without saltation, but may dominate the entrainment process in many arid and semi-arid environments. So it may also be responsible for large magnitude dust storms. Playa and alluvial fan landforms, prior to developing a surface crust, may be the main sources of dust storms in Qinghai Province.

Since 1960, the steppe regions of North Africa have been subject to an increasing desertification, including the degradation of traditional pastures. The initially dominant species (Artemisia herba-alba, Lygeum spartum and Stipa tenacissima) declined and were progressively replaced by other species (Atractylis serratuloides and Salsola vermiculata) that are more tolerant to the new conditions. It is not clear whether these changes are due to anthropogenic reasons or climatic determinism. We have carried out a statistical analysis of the climate to detect putative rainfall changes during the 20^th century in the Algerian steppes based on data from 9 meteorological stations, including 2 Saharan stations (El Oued and Touggourt), 3 pre-Saharan stations (Biskra, Laghouat and Ain Sefra) and 4 steppe stations (Djelfa, Saida, Méchéria and El-Bayadh) located in the arid high plains, which represent the bioclimate diversities of the region. Previous studies suggested that significant rainfall changes for the 20^th century only had records in the south of the Oran region. Most of the studies, however, looked at restricted territories over limited periods, and did not integrate the rainiest period 2004-2014. Our work is designed to integrate all the longest time series of meteorological data available for the steppe regions of Algeria. Our results confirm the spatial rainfall distribution (significant rainfall changes only recorded in the southwestern region) evidenced by previous studies, and reveal a decreasing rainfall gradient from northeastern to southwestern Algeria. Moreover, the results reveal a trend of significant decrease of rainfall in the southern Oran region, marked by two drought periods in 1980-1985 and 1999-2003. However, with the exception of the southwestern region, rainfall overall has not declined since the beginning of the 20^th century. While less marked in other regions, the drought appear to have affected all territories of the Algerian steppe. Consequently, our study implies that the climate was not a leading influence in the on-going degradation of the vegetation cover of steppe landscapes. Such a vegetation evolution thus appears to be have been determined more by human activities than by climate forcing.

Climate change is shifting the amount and frequency of precipitation in many regions, which is expected to affect seedling recruitment across ecosystems. However, the combined effects of precipitation amount and frequency on seedling recruitment remain largely unknown. An understanding of the effects of precipitation amount and frequency and their interaction on seedling emergence and growth of typical desert plants is vital for managing populations of desert plants. We conducted two experiments to study the effects of variation in precipitation on Reaumuria soongarica (Pall.) Maxim. First, greenhouse experiments were conducted to examine the effects of three precipitation amount treatments (ambient, +30%, and -30%) and two precipitation frequency treatments (ambient and -50%) on seedling emergence. Second, the morphological responses of R. soongarica to changes in precipitation amount and frequency were tested in a controlled field experiment. Stage-specific changes in growth were monitored by sampling in different growth seasons. Our results showed that precipitation amount significantly affected germination, seedling emergence, and growth of R. soongarica, and there was a larger effect with decreased precipitation frequency compared with ambient. Germination and seedling emergence increased as precipitation increased under the same frequency of precipitation. The highest emergence percentage was obtained with a 30% increase in precipitation amount and a 50% reduction in precipitation frequency. Compared with ambient precipitation, a 30% increase in precipitation amount increased above- and below-ground biomass accumulation of R. soongarica during the growth season. A decrease of 30% in precipitation amount also resulted in an increase in below-ground biomass and root/shoot ratio in the early stages of the growth season, however, above- and below-ground biomass showed the opposite results at the end of the growth season, with larger effects on above-ground than below-ground biomass under decreased precipitation frequency. When precipitation frequency decreased by 50%, values of all growth traits increased for a given amount of precipitation. We concluded that precipitation frequency may be as important as precipitation amount to seedling emergence and growth of R. soongarica, and that understanding the effects of precipitation variability on seedling recruitment requires the integration of both precipitation amount and frequency. In particular, the combination of a 30% increase in precipitation amount and 50% reduction in precipitation frequency increased the emergence and growth of seedlings, suggesting that alteration of amount and frequency of precipitation caused by climate change may have significant effects on seedling recruitment of R. soongarica.

Grazing is a main human activity in the grasslands of Xinjiang, China. It is vital to identify the effects of grazing on the sustainable utilization of local grasslands. However, the effects of grazing on net primary productivity (NPP), evapotranspiration (ET)and water use efficiency (WUE) in this region remain unclear. Using the spatial Biome-BGC grazing model, we explored the effects of grazing on NPP, ET and WUE across the different regions and grassland types in Xinjiang during 1979-2012. NPP, ET and WUEunder the grazed scenario were generally lower than those under the ungrazed scenario, and the differences showed increasing trends over time.The decreases in NPP, ET and WUE varied significantly among the regions and grassland types. NPP decreasedas follows: among the regions, Northern Xinjiang(16.60 g C/(m²?a)), Tianshan Mountains (15.94g C/(m²?a)) and Southern Xinjiang (-3.54g C/(m²?a)); andamong the grassland types, typical grasslands (25.70g C/(m²?a)), swamp meadows (25.26g C/(m²?a)),mid-mountain meadows (23.39 g C/(m²?a)), alpine meadows (6.33g C/(m²?a)), desert grasslands (5.82g C/(m²?a)) and saline meadows(2.90g C/(m²?a)). ET decreasedas follows: among the regions, Tianshan Mountains (28.95 mm/a), Northern Xinjiang (8.11mm/a) and Southern Xinjiang (7.57 mm/a);and among thegrassland types, mid-mountain meadows (29.30 mm/a), swamp meadows (25.07mm/a), typical grasslands (24.56mm/a), alpine meadows (20.69mm/a), desert grasslands (11.06mm/a) and saline meadows (3.44mm/a). WUE decreasedas follows: among the regions,Northern Xinjiang (0.053g C/kg H₂O), Tianshan Mountains (0.034 g C/kg H₂O) and Southern Xinjiang (0.012 g C/kg H₂O);and among the grassland types, typical grasslands (0.0609 g C/kg H₂O), swamp meadows (0.0548 g C/kg H₂O), mid-mountain meadows (0.0501 g C/kg H₂O), desert grasslands (0.0172 g C/kg H₂O), alpine meadows (0.0121 g C/kg H₂O) and saline meadows (0.0067 g C/kg H₂O). In general, the decreases in NPP and WUE weremore significant in the regions with relatively high levels of vegetation growth because of the high grazing intensity in these regions. The decreases in ET weresignificant in mountainous areas due to theterrain and high grazing intensity.

In order to explore the effects of grazing frequency on functional traits and to test whether Stipa gandis has compensatory photosynthesis during the frequent grazing period, we investigated morphological traits, biomass allocation, photosynthetic traits, and chlorophyll fluorescence parameters of the species in Inner Mongolia, China. The grazing frequency treatments included fencing (T₀), grazing in May and July (T₁, i.e., two months per year) and grazing from May to September (T₂, i.e., continuous five months per year). Results indicate that T₁ and T₂ treatments did not affect individual biomass, but T₂ treatment negatively affected individual size, i.e., plant height, stem length, and leaf length. Physiological traits of S. grandis were significantly affected by grazing, year, and their interaction. In July 2014 (i.e., dry environment and low relative humidity), the photosynthetic rate, transpiration rate and water use efficiency were highest under T₂ treatment, which was caused by the increase in stomatal conductance. However, in July 2015 (i.e., wet environment and high relative humidity), the photosynthetic rate and water use efficiency were higher under T₁ and T₂ treatments, which were caused by the increase in actual quantum efficiency and stomatal conductance. Our results implied that under frequent grazing treatment, S. grandis had small height and efficient compensatory photosynthesis, which promoted its resistance to severe grazing.

Soil organic carbon (SOC) and soil total nitrogen (STN) in arid regions are important components of global C and the N cycles, and their response to climate change will have important implications for both ecosystem processes and global climate feedbacks. Grassland ecosystems of Funyun County in the southern foot of the Altay Mountains are characterized by complex topography, suggesting large variability in the spatial distribution of SOC and STN. However, there has been little investigation of SOC and STN on grasslands in arid regions with a mountain-basin structure. Therefore, we investigated the characteristics of SOC and STN in different grassland types in a mountain-basin system at the southern foot of the Altai Mountains, north of the Junggar Basin in China, and explored their potential influencing factors and relationships with meteorological factors and soil properties. We found that the concentrations and storages of SOC and STN varied significantly with grassland type, and showed a decreasing trend along a decreasing elevation gradient in alpine meadow, mountain meadow, temperate typical steppe, temperate steppe desert, and temperate steppe desert. In addition, the SOC and STN concentrations decreased with depth, except in the temperate desert steppe. According to Pearson's correlation values and redundancy analysis, the mean annual precipitation, soil moisture content and soil available N concentration were significantly positively correlated with the SOC and STN concentrations. In contrast, the mean annual temperature, pH, and soil bulk density were significantly and negatively correlated with the SOC and STN concentrations. The mean annual precipitation and mean annual temperature were the primary factors related to the SOC and STN concentrations. The distributions of the SOC and STN concentrations were highly regulated by the elevation-induced differences in meteorological factors. Mean annual precipitation and mean annual temperature together explained 97.85% and 98.38% of the overall variations in the SOC and STN concentrations, respectively, at soil depth of 0-40 cm, with precipitation making the greatest contribution. Our results provide a basis for estimating and predicting SOC and STN concentrations in grasslands in arid regions with a mountain-basin structure.

Understanding solute transport behaviors of deep soil profile in the Loess Plateau is helpful for ecological construction and agricultural production improvement. In this study, solute transport processes of a deep soil profile were measured by a conservative tracer experiment using 25 undisturbed soil cores (20 cm long and 7 cm diameter for each) continuously sampled from the surface downward to the depth of 500 cm in the Loess Plateau of China. The solute transport breakthrough curves (BTCs) were analyzed in terms of the convection-dispersion equation (CDE) and the mobile-immobile model (MIM). Average pore-water velocity and dispersion coefficient (or effective dispersion coefficient) were calculated using the CDE and MIM. Basic soil properties and water infiltration parameters were also determined to explore their influence on the solute transport parameters. Both pore-water velocity and dispersion coefficient (or effective dispersion coefficient) generally decreased with increasing depth, and the dispersivity fluctuated along the soil profile. There was a good linear correlation between log-transformed pore-water velocity and dispersion coefficient, with a slope of about 1.0 and an average dispersivity of 0.25 for the entire soil profile. Generally speaking, the soil was more homogeneous along the soil profile. Our results also show that hydrodynamic dispersion is the dominant mechanism of solute transport of loess soils in the study area.

Effects of sludge utilization on the mobility and phytoavailability of heavy metals in soil-plant systems have attracted broad attention in recent years. In this study, we analyzed the effects of municipal sludge compost (MSC) on the solubility and plant uptake of Cd, Ni, Cu, Zn and Pb in a soil-potato system to explore the mobility, potato plant uptake and enrichment of these five heavy metals in sierozem soils amended with MSC through a potato cultivation trial in Lanzhou University of China in 2014. Ridge regression analysis was conducted to investigate the phytoavailability of heavy metals in amended soils. Furthermore, CaCl₂, CH₃COONH₄, CH₃COOH, diethylene triamine pentacetic acid (DTPA) and ethylene diamine tetraacetic acid (EDTA) were used to extract the labile fraction of heavy metals from the amended soils. The results show that the MSC could not only improve the fertility but also increase the dissolved organic carbon (DOC) content of sierozem soils. The total concentrations and labile fraction proportions of heavy metals increase with increasing MSC percentage in sierozem soils. In amended soils, Cd has the highest solubility and mobility while Ni has the lowest solubility and mobility among the five heavy metals. The MSC increases the concentrations of heavy metals in the root, stem, peel and tuber of the potato plant, with the concentrations being much higher in the stem and root than in the peel and tuber. Among the five heavy metals, the bioconcentration factor value of Cd is the highest, while that of Ni is the lowest. The complexing agent (DTPA and EDTA) extractable fractions of heavy metals are the highest in terms of phytoavailability. Soil properties (including organic matter, pH and DOC) have important impacts on the phytoavailability of heavy metals. Our results suggest that in soil-potato systems, although the MSC may improve soil fertility, it can also increase the risk of soils exposed to heavy metals.

To investigate the environmental quality of the urban surface soil inYinchuan City, the capital of Ningxia Hui Autonomous Region (Ningxia), China, wesampled surface soil and measured the concentrations of 8 toxic metals (Pb, Cr, Cu, Zn, Co, Bi, Ni and V) using X-ray fluorescence spectrometry. The enrichment characteristics and sources of these toxic metals in the soil were analyzed by the enrichment factor (EF) and multivariate statistical analysis. The results showed that the mean concentrations of these toxic metals in the soil samples were 25.0, 109.1, 16.8, 26.0, 37.2, 2.7, 25.3 and 59.9 mg/kg for Pb, Cr, Cu, Zn, Co, Bi, Niand V, respectively, which were 1.2, 1.8, 0.8, 0.4, 3.2, 8.7, 0.7 and 0.8 times of the corresponding background values of Ningxia soil, respectively. The variations of Pb, Zn, Co, Biand Ni concentrations in the surface soil of Yinchuan were larger than those of the other metals.Our results also showed that the toxic metals investigated in the soil had different enrichment levels. Both Co and Bi were significantly enriched, whereas Cr was only moderatelyenriched in the soil. There was a deficiency or minimal enrichment of the other toxic metals in the soil. Source analysis results based on the concentration, enrichment characteristics and multivariate statistical analysis indicated that Cr, V and Ni originated from a combination of fossil fuel combustion, traffic pollution and natural occurrence.Pb, Cu and Zn were predominantly derived from natural and traffic sources, while Co and Bi primarily originated from construction sources.