Timely and accurate crop identification and mapping are of great significance for crop yield estimation, disaster warning, and food security. Early-season crop identification places higher demands on the quality and mining of time-series information than post-season mapping. In recent years, great strides have been made in the development of deep-learning algorithms, and the emergence of Sentinel-2 data with a higher temporal resolution has provided new opportunities for early-season crop identification. In this study, we aimed to fully exploit the potential of deep-learning algorithms and time-series Sentinel-2 data for early-season crop identification and early-season crop mapping. In this study, four classifiers, i.e., two deep-learning algorithms (one-dimensional convolutional networks and long and short-term memory networks) and two shallow machine-learning algorithms (a random forest algorithm and a support vector machine), were trained using early-season Sentinel-2 images and field samples collected in 2019. Then, these algorithms were applied to images and field samples for 2020 in the Shiyang River Basin. Twelve scenarios with different classifiers and time intervals were compared to determine the optimal combination for the earliest crop identification. The results show that: (1) the two deep-learning algorithms outperformed the two shallow machine-learning algorithms in early-season crop identification; (2) the combination of a one-dimensional convolutional network and 5-day interval time-series Sentinel-2 data outperformed the other schemes in obtaining the early-season crop identification time and achieving early mapping; and (3) the early-season crop identification mapping time in the Shiyang River Basin was identified as the end of July, and the overall classification accuracy reached 0.83. In addition, the early identification time for each crop was as follows: the wheat was in the flowering stage (mid-late June); the alfalfa was in the first harvest (mid-late June); the corn was in the early tassel stage (mid-July); the fennel and sunflower were in the flowering stage (late July); and the melons were in the fruiting stage (around late July). This study demonstrates the potential of using Sentinel-2 time-series data and deep-learning algorithms to achieve early-season crop identification, and this method is expected to provide new solutions and ideas for addressing early-season crop identification monitoring.

The estimation of water requirements constitutes a critical prerequisite for delineating water scarcity hotspots and mitigating intersectoral competition, particularly in endorheic basins in arid or semi-arid regions where hydrological closure exacerbates resource allocation conflicts. Under conditions of water scarcity, water supplied locally by precipitation and shallow groundwater bodies should be taken into account to estimate the net water requirements to be met with water conveyed from off-site sources. This concept is embodied in the distinction of blue ET (BET) and green ET (GET). In this study, the Budyko hypothesis (BH) method was optimized to partition the total ET into GET and BET during 2001–2018 in the Heihe River Basin. In this region, a better knowledge of net water requirements is even more important due to water allocation policies which reduced water supply to irrigated lands in the last 15 years. This study proposes a modified BH method based on a new vegetation-specific parameter (${\omega }_v$) which was optimized for different vegetation types using precipitation and actual ET data obtained from remote sensing observations. The results show that the BH method partitioned GET and BET reasonably well, with a percent bias of 23.8% and 37.4% and a root mean square error of 84.8 mm/a and 113.6 mm/a, respectively, when compared with reported data, which are superior to that of the precipitation deficit and soil water balance methods. A sensitivity experiment showed that the BH method exhibits a low sensitivity to uncertainties of input data. The results documented differences in the contribution of GET and BET to total ET across different land cover types in the Heihe River Basin. As expected, rainfed forest and grassland ecosystems are predominantly governed by GET, with 81.3% and 87.2% of total ET, respectively. In contrast, croplands and shrublands are primarily regulated by BET, with contributions of 61.5% and 84.3% to total ET. The improved BH method developed in this study paves the way for further analyses of the net water requirements in arid and semi-arid regions.

Evaluating the performance of irrigation water use is essential for efficient and sustainable water resource management. However, existing approaches often lack systematic quantification of irrigation water consumption and fail to differentiate between the use of precipitation and anthropogenic appropriation of water flows. Building on the green–blue water concept, consumptive water use, assumed equal to actual evapotranspiration (ET_a), was partitioned into green ET (GET) and blue ET (BET) using remote sensing data and the Budyko hypothesis. A novel BET metric of consumptive irrigation water use was developed and applied to the irrigated lands in northwest China to evaluate the performance of irrigation from 2001 to 2021. The results showed that in terms of total available water resources (precipitation + gross irrigation water (GIW)) compared to irrigation water demand, estimated as reference evapotranspiration (ET₀), Ningxia has sufficient water supply to meet irrigation demand, while the Hexi Corridor faces increasing risks of unsustainable water use. The Hetao irrigation scheme has shifted from a fragile supply–demand balance to a situation where water demand far exceeds availability. In Xinjiang, the balance between water supply and demand is tight. Furthermore, when considering the available water (GIW) relative to the net irrigation water demand (ET₀-GET), the Hexi Corridor faces significant water deficits, and Ningxia and Xinjiang are close to meeting local irrigation water demands by relying on current water availability and efficient irrigation practices. It is noteworthy that the BET remains lower than the GIW in northwest China (excluding the Hexi Corridor in recent years). The ratio of the BET to GIW is an estimate of the scheme irrigation efficiency, which was equal to 0.54 for all irrigation schemes taken together. In addition, the irrigation water use efficiency, estimated as the ratio of BET to net irrigation water, was evaluated in detail, and it was found that in the last 10 years the irrigation water use efficiency improved in Ningxia, the Hetao irrigation scheme, and Xinjiang. However, the Hexi Corridor continues to face severe net irrigation water deficits, suggesting the likelihood of groundwater use to sustain irrigated agriculture. BET innovatively separates consumptive use of precipitation (green water) and consumptive use of irrigation (blue water), a critical advancement beyond conventional approaches’ estimates that merge these distinct hydrological components to help quantifying water use efficiency.