Prediction of crop yield, water consumption and water use efficiency with a SVAT-crop growth model using remotely sensed data on the North China Plain

County-Level Irrigation Water Demand Estimation Using Machine Learning: Case Study of California

Irrigated agriculture is the largest consumer of freshwater globally. Despite the clarity of influential factors and deriving forces, estimation of the volumetric irrigation demand using biophysical models is prohibitively difficult. Data-driven models have proven their ability to predict geophysical and hydrological phenomena with only a handful of influential input variables; however, the lack of reliable input data in most agricultural regions of the world hinders the effectiveness of these approaches. Attempting to estimate the irrigation water demand, we first analyze the correlation of potential influencing variables with irrigation water. We develop machine learning models to predict California’s annual, county-level irrigation water demand based on the statistical analysis findings over an 18-year time span. Input variables are different combinations of deriving meteorological forces, geographical characteristics, cropped area, and crop category. After testing various regression machine learning approaches, the result shows that Gaussian process regression produces the best results. Our findings suggest that irrigated cropped area, air temperature, and vapor pressure deficit are the most significant variables in predicting irrigation water demand. This research also shows that Gaussian process regression can predict irrigation water demand with high accuracy (R2 higher than 0.97 and RMSE as low as 0.06 km3) with different input variable combinations. An accurate estimation of irrigation water use of various crop categories and areas can assist decision-making processes and improve water management strategies. The proposed model can help water policy makers evaluate climatological and agricultural scenarios and hence be used as a decision support tool for agricultural water management at a regional scale.

Prediction of Maize Yield at the City Level in China Using Multi-Source Data

Maize is a widely grown crop in China, and the relationships between agroclimatic parameters and maize yield are complicated, hence, accurate and timely yield prediction is challenging. Here, climate, satellite data, and meteorological indices were integrated to predict maize yield at the city-level in China from 2000 to 2015 using four machine learning approaches, e.g., cubist, random forest (RF), extreme gradient boosting (Xgboost), and support vector machine (SVM). The climate variables included the diffuse flux of photosynthetic active radiation (PDf), the diffuse flux of shortwave radiation (SDf), the direct flux of shortwave radiation (SDr), minimum temperature (Tmn), potential evapotranspiration (Pet), vapor pressure deficit (Vpd), vapor pressure (Vap), and wet day frequency (Wet). Satellite data, including the enhanced vegetation index (EVI), normalized difference vegetation index (NDVI), and adjusted vegetation index (SAVI) from the Moderate Resolution Imaging Spectroradiometer (MODIS), were used. Meteorological indices, including growing degree day (GDD), extreme degree day (EDD), and the Standardized Precipitation Evapotranspiration Index (SPEI), were used. The results showed that integrating all climate, satellite data, and meteorological indices could achieve the highest accuracy. The highest estimated correlation coefficient (R) values for the cubist, RF, SVM, and Xgboost methods were 0.828, 0.806, 0.742, and 0.758, respectively. The climate, satellite data, or meteorological indices inputs from all growth stages were essential for maize yield prediction, especially in late growth stages. R improved by about 0.126, 0.117, and 0.143 by adding climate data from the early, peak, and late-period to satellite data and meteorological indices from all stages via the four machine learning algorithms, respectively. R increased by 0.016, 0.016, and 0.017 when adding satellite data from the early, peak, and late stages to climate data and meteorological indices from all stages, respectively. R increased by 0.003, 0.032, and 0.042 when adding meteorological indices from the early, peak, and late stages to climate and satellite data from all stages, respectively. The analysis found that the spatial divergences were large and the R value in Northwest region reached 0.942, 0.904, 0.934, and 0.850 for the Cubist, RF, SVM, and Xgboost, respectively. This study highlights the advantages of using climate, satellite data, and meteorological indices for large-scale maize yield estimation with machine learning algorithms.

Investigating the spatiotemporal differences and influencing factors of green water use efficiency of Yangtze River Economic Belt in China

Combining freshwater consumption and wastewater emissions into a unified analysis framework and utilizing the epsilon-based measure (EBM) model with the characteristics of radial model and non-radial model, this paper evaluates green water use efficiency (GWUE) of 11 provincial-regions in the Yangtze River Economic Belt (YREB) and investigates its spatiotemporal differences during the period 2005–2014, on basis of which the contribution rate of each input-specific green water use inefficiency in the overall green water use efficiency and the potential of freshwater-saving and wastewater emissions reduction are also calculated. The Theil index is used to explore the sources of the provincial gap of green water use inefficiency, and a random-effect panel Tobit model is applied to test the impact of the influencing factors of green water use inefficiency in the YREB. It is found that green water use inefficiency of the YREB is relatively low and regional differences is significant during the sample period, indicating a large potential of water-saving and water pollution reduction, and narrowing BGAP and WGAP of the Upstream is the key for improving green water use inefficiency in the YREB. The panel Tobit regression results show that economic development, technological innovation, water use structure, water resources endowment, environmental regulation and regional differences all play positive/negative effects on green water use inefficiency in the YREB, while these factors’ influencing direction, degree and significance are significantly different. The conclusions of our study can provide considerably valuable information for the YREB to reserve water resources and reduce wastewater emissions.

Prediction of Winter Wheat Yield Based on Multi-Source Data and Machine Learning in China

Wheat is one of the main crops in China, and crop yield prediction is important for regional trade and national food security. There are increasing concerns with respect to how to integrate multi-source data and employ machine learning techniques to establish a simple, timely, and accurate crop yield prediction model at an administrative unit. Many previous studies were mainly focused on the whole crop growth period through expensive manual surveys, remote sensing, or climate data. However, the effect of selecting different time window on yield prediction was still unknown. Thus, we separated the whole growth period into four time windows and assessed their corresponding predictive ability by taking the major winter wheat production regions of China as an example in the study. Firstly we developed a modeling framework to integrate climate data, remote sensing data and soil data to predict winter wheat yield based on the Google Earth Engine (GEE) platform. The results show that the models can accurately predict yield 1~2 months before the harvesting dates at the county level in China with an R2 > 0.75 and yield error less than 10%. Support vector machine (SVM), Gaussian process regression (GPR), and random forest (RF) represent the top three best methods for predicting yields among the eight typical machine learning models tested in this study. In addition, we also found that different agricultural zones and temporal training settings affect prediction accuracy. The three models perform better as more winter wheat growing season information becomes available. Our findings highlight a potentially powerful tool to predict yield using multiple-source data and machine learning in other regions and for crops.

Contributions of climate change and vegetation greening to evapotranspiration trend in a typical hilly-gully basin on the Loess Plateau, China

Significant increases in vegetation cover on the Loess Plateau since the early 2000s have been well documented. However, the relevant hydrological effects are still unclear. Here, we investigated the changes in actual evapotranspiration (ET_a) from 2000 to 2016 and related them to climate change and vegetation greening in Yanhe River basin (YRB), a typical hilly-gully basin on the Loess Plateau, by using the remote-sensing based VIP model. Results showed that the annual ET_a in the YRB increased significantly with a trend of 3.45mmyr^-1 (p<0.01) and changes of ET_a in summer months dominated the annual trend. Partial correlation analysis suggested that vegetation greening was the dominant driving factor of ET_a inter-annual variations in 56% area of YRB. Model simulation experiments illustrated that relative contributions of NDVI, precipitation, and potential evapotranspiration (ET_p) to the ET_a trend were 93.0%, 18.1%, and -7.4%, respectively. Vegetation greening, which is closely related to the Grain for Green (GFG) afforestation, was the main driver to the long-term tendency of water consumption in the YRB. This study highlights potential water demanding conflicts between the socio-economic system and the natural ecosystem on the Loess Plateau due to the rapid vegetation expansion in this water-limited area.

Effects of water stress on water use efficiency of irrigated and rainfed wheat in the Loess Plateau, China

The Loess Plateau, the largest arid and semi-arid zone in China, has been confronted with more severe water resource pressure and a growing demand for food production under global changes. For developing sustainable agriculture in this region, it is critical to learn spatiotemporal variations in water use efficiency (WUE) of main crops (e.g. winter wheat in this region) under various water management practices. In this study, we classified irrigated and rainfed wheat areas based on MODIS data, and calculated the winter wheat yield by using an improved light use efficiency model. The actual evapotranspiration (ETa) of winter wheat and the evapotranspiration drought index (EDI) were also investigated. Then we mainly examined the synergistic relationship between crop yield, ETa, and WUE, and analyzed the variations in WUE of irrigated and rainfed wheat under water stress during the 2010-2011 growing season. The results suggested that winter wheat in the Loess Plateau was primarily dominated by rainfed wheat. The average yield of irrigated wheat was 3928.4 kg/ha, 22.2% more than that of rainfed wheat. High spatial heterogeneities of harvest index (HI) and maximum light use efficiency (ε_max) were found in the Loess Plateau. The ETa of irrigated wheat was 10.2% more than that of rainfed wheat. The ratio of irrigated and rainfed wheat under no water stress was 31.55% and 17.16%, respectively. With increasing water stress, the WUE of rainfed wheat decreased more quickly than that of irrigated wheat. The WUE variations in winter wheat under water stress depended strongly on the synergistic effects of two WUE components (crop yield and ETa) and their response to environmental conditions as well as water management practices (irrigated or rainfed). Our findings enhance our current understanding of the variations in WUE as affected by water stress under various water use conditions in arid and semi-arid areas.

Nationwide Projection of Rice Yield Using a Crop Model Integrated with Geostationary Satellite Imagery: A Case Study in South Korea

The Geostationary Ocean Color Imager (GOCI) of the Communication, Ocean, and Meteorological Satellite (COMS) increases the chance of acquiring images with greater clarity eight times a day and is equipped with spectral bands suitable for monitoring crop yield in the national scale with a spatial resolution of 500 m. The objectives of this study were to classify nationwide paddy fields and to project rice (Oryza sativa) yield and production using the grid-based GRAMI-rice model and GOCI satellite products over South Korea from 2011 to 2014. Solar insolation and temperatures were obtained from COMS and the Korea local analysis and prediction systems for model inputs, respectively. The paddy fields and transplanting dates were estimated by using Moderate Resolution Imaging Spectroradiometer (MODIS) reflectance and land cover products. The crop model was calibrated using observed yield data in 11 counties and was applied to 62 counties in South Korea. The overall accuracies of the estimated paddy fields using MODIS data ranged from 89.5% to 90.2%. The simulated rice yields statistically agreed with the observed yields with mean errors of −0.07 to +0.10 ton ha−1, root-mean-square errors of 0.219 to 0.451 ton ha−1, and Nash–Sutcliffe efficiencies of 0.241 to 0.733 in four years, respectively. According to paired t-tests (α = 0.05), the simulated and observed rice yields were not significantly different. These results demonstrate the possible development of a crop information delivery system that can classify land cover, simulate crop yield, and monitor regional crop production on a national scale.

Closing water productivity gaps to achieve food and water security for a global maize supply

To achieve food and water security, it is as important to close the water productivity (WP) gap (which was defined as the difference between the maximum attainable WP and the currently achieved WP at the field scale) as it is to close yield gaps. However, few studies have provided quantitative estimates of existing WP gaps and constraining factors for global maize production. Using a meta-analysis of 473 published studies covering 31 countries and 5,553 observations (932 site-years), we found the global average WP value for irrigated maize was 18.6 kg ha⁻¹ mm⁻¹. These WPs varied by factors such as seasonal precipitation, irrigation regimes, soil organic matter and soil pH. In current production systems, there exists a huge scope for improvement in maize WP, but the reported field experiments achieved only 20–46% of potential WP across all countries. Considering the future, raising WP to 85% of potential WP by 2050, a 100% increase in maize production could be achieved with 20% less planted area and 21% less water consumption than in 2005. Closing the WP gap may be critical to ensuring food security and achieving sustainable global agriculture.

Re-Examining Regional Total-Factor Water Efficiency and Its Determinants in China: A Parametric Distance Function Approach

It is accepted that improving water efficiency is a key task for China in achieving water sustainability, as the knowledge of water efficiency and its determinants can provide critical information for water policy formulation. To this end, this paper presents a parametric frontier approach to analyze water efficiency performance and its influencing factors in one step. The proposed approach first introduces the Shephard water distance function to construct total-factor water efficiency (TFWE) index and then adopts the stochastic frontier analysis (SFA) technique to compute the index and its determinants. A case study of regions in China from 2000 to 2015 is presented. The main findings are summarized as follows: (1) Both the overall China and most of the regions still have room for improvement in water efficiency. SFA and data envelopment analysis (DEA) might lead to different results in benchmarking water efficiency. Moreover, SFA has higher discriminating power than DEA in this regard. (2) There exists significant disparity of water efficiency among the regions of China, and the difference in TFWE takes on a U-shaped evolution trend, which first decreases in a fluctuation way and then increases monotonically. (3) Factors like industrial structure, import and export trade, environmental regulation and urbanization level have a positive impact on water efficiency, while resource endowment and economic level exhibit negative and nonlinear effects, respectively. Finally, several policy recommendations are made to improve water efficiency levels and promote water sustainability.

Forecasting of Industrial Water Demand Using Case-Based Reasoning—A Case Study in Zhangye City, China

Forecasting the industrial water demand accurately is crucial for sustainable water resource management. This study investigates industrial water demand forecasting by case-based reasoning (CBR) in an arid area, with a case study of Zhangye, China. We constructed a case base with 420 original cases of 28 cities in China, extracted six attributes of the industrial water demand, and employed a back propagation neural network (BPN) to weight each attribute, as well as the grey incidence analysis (GIA) to calculate the similarities between target case and original cases. The forecasting values were calculated by weighted similarities. The results show that the industrial water demand of Zhangye in 2030, which is the t arget case, will reach 11.9 million tons. There are 10 original cases which have relatively high similarities to the target case. Furthermore, the case of Yinchuan, 2010, has the largest similarity, followed by Yinchuan, 2009, and Urumqi, 2009. We also made a comparison experiment in which case-based reasoning is more accurate than the grey forecast model and BPN in water demand forecasting. It is expected that the results of this study will provide references to water resources management and planning.

Winter wheat water productivity evaluated by the developed remote sensing evapotranspiration model in Hebei Plain, China

Agricultural water is the main reason for the rapid decline of the NCP groundwater levels. It is of vital importance for the NCP sustainable agricultural development to master the ETa and its CWP. In this paper, the EBEM model was developed according to the theory of energy balance. From 2001 to 2006, the winter wheat ETa and CWP were estimated, and the spatial and temporal variations and their influencing factors were studied in the Hebei Plain. The results indicate that the EBEM model performed well by applying MODIS data to estimate the daily net radiation and ETa. For the daytime net radiation, the relative error between the estimation and the measurement amounted to 8.2% and the SEE was 0.82 MJ m(-2)/day. The average ETa deviation between the estimates and the measures amounted to 0.86 mm daily, and the SE was 1.2 mm. The spatial variations indicated that the major distribution of ETa ranged from 350 to 450 mm, which trended downward within the study area from west to east. In the study period, the winter wheat CWP was mainly distributed between 0.29 and 1.67 kg/m(3). In space, the CWP was higher in the west than in the east.

Quantifying climate and management effects on regional crop yield and nitrogen leaching in the north china plain

Better water and nitrogen (N) management requires better understanding of soil water and N balances and their effects on crop yield under various climate and soil conditions. In this study, the calibrated Root Zone Water Quality Model (RZWQM2) was used to assess crop yield and N leaching under current and alternative management practices in a double-cropped wheat ( L.) and maize ( L.) system under long-term weather conditions (1970-2009) for dominant soil types at 15 locations in the North China Plain. The results provided quantitative long-term variation of deep seepage and N leaching at these locations, which strengthened the existing qualitative knowledge for site-specific management of water and N. In general, the current management practices showed high residual soil N and N leaching in the region, with the amounts varying between crops and from location to location and from year to year. Seasonal rainfall explained 39 to 84% of the variability in N leaching (1970-2009) in maize across locations, while for wheat, its relationship with N leaching was significant ( < 0.01) only at five locations. When N and/or irrigation inputs were reduced to 40 to 80% of their current levels, N leaching generally responded more to N rate than to irrigation, while the reverse was true for crop yield at most locations. Matching N input with crop requirements under limited water conditions helped achieve lower N leaching without considerable soil N accumulation. Based on the long-term simulation results and water resources availability in the region, it is recommended to irrigate at 60 to 80% of the current water levels and fertilize only at 40 to 60% of the current N rate to minimizing N leaching without compromising crop yield.

Remotely sensed rice yield prediction using multi-temporal NDVI data derived from NOAA's-AVHRR

Grain-yield prediction using remotely sensed data have been intensively studied in wheat and maize, but such information is limited in rice, barley, oats and soybeans. The present study proposes a new framework for rice-yield prediction, which eliminates the influence of the technology development, fertilizer application, and management improvement and can be used for the development and implementation of provincial rice-yield predictions. The technique requires the collection of remotely sensed data over an adequate time frame and a corresponding record of the region's crop yields. Longer normalized-difference-vegetation-index (NDVI) time series are preferable to shorter ones for the purposes of rice-yield prediction because the well-contrasted seasons in a longer time series provide the opportunity to build regression models with a wide application range. A regression analysis of the yield versus the year indicated an annual gain in the rice yield of 50 to 128 kg ha⁻¹. Stepwise regression models for the remotely sensed rice-yield predictions have been developed for five typical rice-growing provinces in China. The prediction models for the remotely sensed rice yield indicated that the influences of the NDVIs on the rice yield were always positive. The association between the predicted and observed rice yields was highly significant without obvious outliers from 1982 to 2004. Independent validation found that the overall relative error is approximately 5.82%, and a majority of the relative errors were less than 5% in 2005 and 2006, depending on the study area. The proposed models can be used in an operational context to predict rice yields at the provincial level in China. The methodologies described in the present paper can be applied to any crop for which a sufficient time series of NDVI data and the corresponding historical yield information are available, as long as the historical yield increases significantly.

Two Improvements of an Operational Two-Layer Model for Terrestrial Surface Heat Flux Retrieval

In order to make the prediction of land surface heat fluxes more robust, two improvements were made to an operational two-layer model proposed previously by Zhang. These improvements are: 1) a surface energy balance method is used to determine the theoretical boundary lines (namely ‘true wet/cool edge’ and ‘true dry/warm edge’ in the trapezoid) in the scatter plot for the surface temperature versus the fractional vegetation cover in mixed pixels; 2) a new assumption that the slope of the T_m – f curves is mainly controlled by soil water content is introduced. The variables required by the improved method include near surface vapor pressure, air temperature, surface resistance, aerodynamic resistance, fractional vegetation cover, surface temperature and net radiation. The model predictions from the improved model were assessed in this study by in situ measurements, which show that the total latent heat flux from the soil and vegetation are in close agreement with the in situ measurement with an RMSE (Root Mean Square Error) ranging from 30 w/m²∼50 w/m², which is consistent with the site scale measurement of latent heat flux. Because soil evaporation and vegetation transpiration are not measured separately from the field site, in situ measured CO₂ flux is used to examine the modeled λE_veg. Similar trends of seasonal variations of vegetation were found for the canopy transpiration retrievals and in situ CO₂ flux measurements. The above differences are mainly caused by 1) the scale disparity between the field measurement and the MODIS observation; 2) the non-closure problem of the surface energy balance from the surface fluxes observations themselves. The improved method was successfully used to predict the component surface heat fluxes from the soil and vegetation and it provides a promising approach to study the canopy transpiration and the soil evaporation quantitatively during the rapid growing season of winter wheat in northern China.