Soil moisture content (SMC) is an important factor that affects agricultural development in arid regions. Compared with the space-borne remote sensing system, the unmanned aerial vehicle (UAV) has been widely used because of its stronger controllability and higher resolution. It also provides a more convenient method for monitoring SMC than normal measurement methods that includes field sampling and oven-drying techniques. However, research based on UAV hyperspectral data has not yet formed a standard procedure in arid regions. Therefore, a universal processing scheme is required. We hypothesized that combining pretreatments of UAV hyperspectral imagery under optimal indices and a set of field observations within a machine learning framework will yield a highly accurate estimate of SMC. Optimal 2D spectral indices act as indispensable variables and allow us to characterize a model’s SMC performance and spatial distribution. For this purpose, we used hyperspectral imagery and a total of 70 topsoil samples (0–10 cm) from the farmland (2.5 × 104 m2) of Fukang City, Xinjiang Uygur AutonomousRegion, China. The random forest (RF) method and extreme learning machine (ELM) were used to estimate the SMC using six methods of pretreatments combined with four optimal spectral indices. The validation accuracy of the estimated method clearly increased compared with that of linear models. The combination of pretreatments and indices by our assessment effectively eliminated the interference and the noises. Comparing two machine learning algorithms showed that the RF models were superior to the ELM models, and the best model was PIR (R2val = 0.907, RMSEP = 1.477, and RPD = 3.396). The SMC map predicted via the best scheme was highly similar to the SMC map measured. We conclude that combining preprocessed spectral indices and machine learning algorithms allows estimation of SMC with high accuracy (R2val = 0.907) via UAV hyperspectral imagery on a regional scale. Ultimately, our program might improve management and conservation strategies for agroecosystem systems in arid regions.

No abstract

The Ebinur Lake watershed is an important ecological barrier for environmental changes in the Junggar Basin in Xinjiang Uygur Autonomous Region (XUAR). Due to the tremendous changes in the underlying surface environment of the watershed in the past few decades, the watershed has become a typical region of ecological degradation. Drought affects the surface dynamics and characterizes the regional dry and wet environments, while the dynamic variation in lakes and vegetation are indicators of dynamic changes in land surfaces. Thus, a quantitative assessment of the response of lakes and vegetation to drought conditions at multiple temporal scales is critical for assessing the potential impacts of regional climate change on terrestrial ecosystems and ecological restoration. The standardized precipitation evapotranspiration index (SPEI), the spectral water index (NDWI) and the normalized difference vegetation index (NDVI) were used to analyse the evolution of drought, the variation in lake surface area and the sustainable variation in vegetation. Furthermore, we quantitatively evaluated the response patterns of vegetation to droughts of multiple temporal scales (1-, 3-, 6-, 12-, 24-month). The conclusions showed that (1) overall, the area of Ebinur Lake experienced drastic fluctuations, and the lake area has decreased significantly since 2003, with a dynamic area of 817.63 km2 in 2003 and 384.60 km2 in 2015, and the lake area had shrank severely. (2) The interannual variation of wet and dry changed alternately during the observation period, and persistent drought events occurred from 2006 to 2010 across the Ebinur Lake watershed. (3) The vegetation area of cultivated land expanded continuously across the watershed, and the grassland degraded severely. (4) The changes in lake surface area are significantly correlated with the inflow water volume (correlation coefficient = 0.64, P < 0.01). (5) The vegetation of different terrestrial ecosystems exhibited heterogeneous responses to multiple temporal scales of drought in different seasons. The percentage was 72.78% of the total area, which showed a correlation between vegetation and drought conditions during the growing season period, and there were more impacts of drought on vegetation, with values as high as 64.33% of the area in summer, than those in other seasons.

X. 2019. Combining UAV-based hyperspectral imagery and machine learning algorithms for soil moisture content monitoring. PeerJ 7:e6926 https://doi.