| Oases are non-zonal geographical views in desert areas relying heavily on water resources.As an integral part of the water circulation,soil water affects the mass transport and the energy flow in soil-plant-atmosphere continuum,and plays a vital role in maintaining the sustainability of dryland ecosystems.Understanding the dynamics of soil moisture is of great significance in regional eco-hydrological modeling,water resources management,and vegetation construction.In order to explore the landscape-scale spatial and temporal evolution of soil water in complex ecosystems,we carried out a case study in a desert oasis in the middle reaches of the Heihe River basin.The data sets were collected in a 0–300 cm soil profile of a ~100 km2 transect across multiple land cover types over a three-year period(2012 and 2014).The variations of soil water in space and time and its influencing factors were investigated using classical statistical analysis,geostatistical analysis,fractal dimensions analysis,and gray correlation analysis.The study focused on the variability of soil water content in different land uses,the difference of temporal stability indexes,the identification of representative locations in soil water monitoring,the relationship between soil water distribution and vegetation pattern,the gradient of soil water in oases along the river bank,and the assessment of soil drought using stochastic simulation.The main results obtained were listed as follows:1.Soil water was unevenly distributed across multiple land cover types by the following order for all soil layers: grassland > cropland > forestland > shrubland > desert.The spatial-temporal variations of soil moisture were depth dependent.Generally,the spatial variability of soil water decreased with soil depth in cropland,forestland,and grassland,while increased with depth in shrubland and desert.Land use type and soil depth worked together to determine the relationship between spatial variability and mean soil water content.The spatial variability decreased with mean soil water content in 60–100 cm and140–200 cm of cropland and in 60–200 cm of forestland,but increased with mean soil water content in 60–100 cm and 100–140 cm of shrubland and in 60–200 cm of desert.2.The spatial distribution of soil water exert control on the vegetation patterns.Soil water contents in all soil layers of the 0–300 cm profile were significantly correlated with vegetation coverage.The relationships between soil water content and vegetation were higher in the layers of 0–100 cm than the layers below 100 cm and reached the highest in the layers of 40–80 cm.The ranges of cross variograms between soil water content and normalized difference vegetation index(NDVI)varied between 4694 m and 6300 m.Land uses had significant effects on the relationship between soil water and vegetation.There existed no significant correlations between soil water and vegetation in cropland,forestland and grassland,but there existed significant correlations between them in 60–300 cm of shrubland and in 0–220 cm of desert.3.Soil moisture was organized with apparent spatial autocorrelation for various soil layers over the entire region.The spatial patterns of soil moisture were generally more complex in upper soils due to short-range variations indicated by the overall increasing spatial correlation lengths and decreasing fractal dimensions with soil depth.The correlation length was negatively related to field mean soil moisture in the surface layer(0–20 cm),but positively related to field mean soil moisture in the 100–140 cm layer.4.The spatial patterns of soil moisture exhibited great similarity over the sampling domain from repeated measurements.The temporal stability of the spatial pattern of soil moisture withered over time and fluctuated with the season.Soil moisture patterns exhibited higher stability between two same seasons than between two different seasons.The local temporal stability of soil moisture,measured by standard deviation of relative difference(SDRD)and mean absolute bias error(MABE),were spatially autocorrelated and had medium spatial variability.SDRD and MABE made great differences in evaluating the temporal stability,including the kriging maps and the representative locations.We found negative correlations between SDRD and MABE in the surface layer of 0–20 cm and positive correlations between them in the layers of 120–140 cm and 180–200 cm.The choice of indices was critical to understanding the developing mechanism of temporal stability.Soil moisture condition,and soil particle composition were noticeable core factors to the temporal stability of soil moisture.We got an accurate estimation of field mean soil moisture for each soil layer using the representative location in the old-oasis croplands.But there is no serviceability to estimate the field mean soil moisture accurately using a single location in the newly-exploited croplands.Hence,the newly-exploited croplands and the old-oasis croplands should be treated differently when monitoring the soil moisture content in the field.5.The mean soil water storage in the 2 m profile of river bank oases was approximately 367 mm.There existed a visible gradient in the distribution of soil water storage from the oasis to the desert,which was strongly influenced by land use types.On average,the soil water storage in the cropland was more than 300% above that in desert.The river channel had direct effects on soil water storage within the distance of 3 km.A time-stable transect was observed for soil water storage between the river bank oasis and the newly cultivated cropland.This transect could help to evaluate the health and development of the oasis ecosystem by assessing the temporal stability of soil water.6.Two hundred realizations described the possible spatial distributions of soil moisture for each month.Averaging the overall realizations provided the expected spatial patterns of soil moisture over the study area.We defined soil drought as a natural event that occurred when the spatial probability of soil moisture not exceeding 0.15 cm3 cm-3 was higher than a critical threshold,and then forecasted the occurrence of soil drought under different regulatory probabilities.The proportion of soil drought ares in each land use declined at different rates with the increasing of probability thresholds.Considering the values of 0.6as the critical threshold of spatial probability,1.3–3.8% of the cultivated land,2.6–5.2%of the forest land,3.2–4.6% of the grassland,and 2.7–7.4% of the shrub land were of soil drought over the measuring period.To prevent the probable damage to the oasis ecosystem,some protection measures should be considered in the SMS locations during the decisionmaking processes.The study demonstrated that land use type and soil mechanical composition were the main influencing factors of soil water distribution.Although soil water was highly variable in space and time,it exhibited strong spatial structure and temporal stability.The temporal stability analysis can be used to optimize the soil water monitoring network and to indicate the developments of desert and oasis ecosystems.Stochastic simulation is an effective tool to assess the risk of soil drought.To prevent land degradation and maintain the sustainability of the agroecosystem,it is imperative to stay on guard against soil moisture scarcity in the oasis.Improving the irrigation program of cropland and increasing the irrigation volume of forestland might be wise moves.These insights will help to understand the ecological and hydrological processes in desert oasis and provide scientific basis for vegetation construction and environmental conservation in arid areas. |
PDF Full Text Request