Identifying long-term winter wheat planting areas using decision tree-derived ensemble learning models and multi-source data

Winter wheat is a staple food crop for most of the world’s population, and the area and spatial distribution of winter wheat are key elements in estimating crop production and ensuring food security. However, winter wheat planting areas contain substantial spatial heterogeneity with mixed pixels for coarse- and moderate-resolution satellite data, leading to large errors in crop acreage estimation. This study has developed a phenology-based approach using moderate-resolution (1 km per pixel) satellite data to estimate sub-pixel planting fractions of winter wheat. Based on unmanned aerial vehicle (UAV) observations, the unique characteristics of winter wheat with high vegetation index values at the heading stage (May) and low values at the ripening stage (June) were investigated. The differences in vegetation index between heading and ripening stages increased with the planting fraction of winter wheat, and therefore the planting fractions were estimated by comparing the NDVI differences of a given pixel with those of predetermined pure winter wheat and non-winter wheat pixels. This approach was evaluated using aerial images and agricultural statistical data in an intensive agricultural region, Shandong Province in North China. The method explained 85% and 60% of the spatial variation in municipal- and county-level statistical data, respectively. More importantly, the predetermined pure winter wheat and non-winter wheat pixels can be automatically identified using MODIS data according to their NDVI differences, which strengthens the potential to use this method at regional and global scales without any field observations as references.

Abstract. Lack of nitrogen leads to a decrease in yield and grain quality in winter wheat plants. Therefore, it is necessary to monitor nitrogen nutrition throughout the period of growth and development of plants, which will quickly assess the need for fertilizing to obtain high yields of quality grain. Therefore, the aim of the study was to establish the possibility of using the normalized difference vegetation index (NDVI) to control the nitrogen content in winter wheat plants in the Stavropol territory. Methods. The work was performed in federal state budgetary scientific institution “North-Caucasian Federal Agricultural Research Centre” at the production of winter crops. Selection of plant samples (sheaf material) was carried out according to the generally accepted method. Repeated – 4x. Determination of the chemical composition of plant organs was carried out by the method of V. T. Kurkaev with co-authors, and the content of chlorophyll – Milaeva and Primak. Results. Since the quality of winter wheat grain directly depends on the nitrogen supply of plants, the relationships between the nitrogen content in winter wheat plants and the values of the vegetation index NDVI were studied. High correlation coefficients between these indicators are obtained. Thus, the average of Rcorr fields.in 2012 it was equal to –0.89, and in 2013 and 2014 –0.82. In addition, due to the dependence of nitrogen content on the amount of chlorophyll, it was possible to analyze the correlation between these indicators and NDVI fields, which showed that a stable relationship (inverse) is observed in the case of the amount of chlorophyll per unit biomass (mg/g), which is estimated on average at –0.79. The interrelation between grain quality and earth remote sensing data is revealed. It is most clearly seen in the case of the maximum and average NDVI for the period from the resumption of spring vegetation to full ripeness of winter wheat. Scientific novelty. For the first time in the conditions of unstable humidification of the Stavropol territory, a high inverse correlation between the vegetation index NDVI and the nitrogen content in winter wheat plants was determined, which on average is estimated by the correlation coefficient equal to –0.84.

Winter wheat is one of the most important crops planted in Beijing suburb. In recent 20 years, winter wheat planted area decreased obviously in Beijing area owning to the urbanization process. This study focuses on the winter wheat planted area transformation monitoring of Beijing suburb from 1992 to 2009 through remote sensing technique. Multi-temporal Landsat- TM images are collected during the winter wheat growth season of 1992，2000 and 2009 and used to analyze the trend and characteristics of winter wheat field variation in Beijing suburb in recent two decades years. The PCA analysis and Tasseled Cap transform technique are adopted in this study for feature classification. The study result shows that the winter wheat planted area in 1992，2000 and 2009 in Beijing is 113671 ha，84322 ha and 61529 ha, respectively. It indicates that winter wheat planting area in Beijing has a significantly decreasing trend and the total reduced area is 52143 ha from 1992 to 2009. Winter wheat planted area is decreased by 29349 ha from 1992 to 2000. Most of reduced wheat fields are transformed into bare land or used for urban land accounting for 42.8% and 39.7%. Others wheat fields are used for greenhouse and water bodies (fish ponds and water fields), accounting for 13.3% and 3%. The winter wheat field decreased by 22794 ha from 2000 to 2009, more than 41.93% of wheat field is turned into bare land. Reduce field for greenhouse land and water bodies (ponds or water fields) are account for21.61% and 7.79%, respectively.

An investigation into the impact of the maximum Normalized Difference Vegetation Index (NDVI) and the maximum surface temperature (Ts) compositing procedures (MaN and MaT respectively) upon retrieved NDVI and Ts values extracted from forested areas located across eight months of cloud screened European AVHRR data is described. NDVI values are found to be significantly higher and generally less variable when they are extracted from MaN rather than from MaT composites and Ts values are found to be significantly higher and generally less variable when they are extracted from MaT rather than from MaN composites. The impact of these differences is illustrated within the context of a European forest/non-forest classification that uses both NDVI and Ts data. Higher potential forest/non-forest classification accuracies are found using NDVI data extracted from the MaN composites and Ts data extracted from the MaT composites than from any other combination of composited data. The findings indicate that inappropriate selection of a compositing procedure may have a significant impact upon the subsequent application of NDVI and/or Ts data.

This research utilized “GaoFen-1” 2 m panchromatic, 8 m and 16 m multispectral satellite data to monitor the planting area of winter wheat. Calculate the correction coefficient by the results of planting area of winter wheat gotten from “GaoFen-1” 16 m multispectral satellite data and micro unmanned aerial vehicle data. Correct the area extraction result of “GaoFen-1” 16 m multispectral satellite data and verify the accuracy of the results got from fusion data of 2 m resolution and 8 m resolution of “GaoFen-1” by micro unmanned aerial vehicle data. Then compare with the area extraction results, extracted from the data of fusion data of 2 m resolution and 8 m resolution of “GaoFen-1” and the corrected 16 m multispectral satellite data, in Jizhou. The study result show that the corrected extraction result of planting area of winter wheat, from “GaoFen-1” 16 m multispectral satellite data, is accurate relatively and could better meet the requirements of the crop dynamic monitoring business operational running.

Precipitation is a major determinant of vegetation production at the regional scale, especially in tropical areas. Recent works, however, have shown that this relationship is weak at the interannual temporal scale, particularly in mesic ecosystems, where limitations other than water constrain vegetation production. We investigated whether this holds true for eastern and southern African savannas, by studying the relationship between precipitation and the normalized difference vegetation index (NDVI) at the interannual time‐scale in 33 protected areas. We also used extreme precipitation and NDVI events to reveal the overlooked influence of precipitation along the mean annual precipitation (MAP) gradient. Only the semi‐arid ecosystems showed significant precipitation–NDVI relationships. We found that maximum NDVI was associated with opposite precipitation conditions along the MAP gradient; maximum NDVI was associated with some of the largest precipitation in semi‐arid sites and with some of the lowest precipitation in mesic sites. Although untested, these results are consistent with the hypothesis of an interaction between water and nutrient limitations along the MAP gradient. Our results extend to African savannas the previous finding that ecosystem sensitivity to annual precipitation decreases with increasing MAP, and highlight that, even in mesic ecosystems, precipitation patterns condition the likelihood of reaching maximum NDVI.

Crop yield forecasting is performed monthly during the growing season by the United States Department of Agriculture’s National Agricultural Statistics Service. The underpinnings are long-established probability surveys reliant on farmers’ feedback in parallel with biophysical measurements. Over the last decade though, satellite imagery from the Moderate Resolution Imaging Spectroradiometer (MODIS) has been used to corroborate the survey information. This is facilitated through the Global Inventory Modeling and Mapping Studies/Global Agricultural Monitoring system, which provides open access to pertinent real-time normalized difference vegetation index (NDVI) data. Hence, two relatively straightforward MODIS-based modeling methods are employed operationally. The first model constitutes mid-season timing based on the maximum peak NDVI value, while the second is reflective of late-season timing by integrating accumulated NDVI over a threshold value. Corn model results nationally show the peak NDVI method provides a R2 of 0.88 and a coefficient of variation (CV) of 3.5%. The accumulated method, using an optimally derived 0.58 NDVI threshold, improves the performance to 0.93 and 2.7%, respectively. Both these models outperform simple trend analysis, which is 0.48 and 7.4%, correspondingly. For soybeans the R2 results of the peak NDVI model are 0.62, and 0.73 for the accumulated using a 0.56 threshold. CVs are 6.8% and 5.7%, respectively. Spring wheat’s R2 performance with the accumulated NDVI model is 0.60 but just 0.40 with peak NDVI. The soybean and spring wheat models perform similarly to trend analysis. Winter wheat and upland cotton show poor model performance, regardless of method. Ultimately, corn yield forecasting derived from MODIS imagery is robust, and there are circumstances when forecasts for soybeans and spring wheat have merit too.

In this study, relationships between normalized difference vegetation index (NDVI) and plant (winter wheat) nitrogen content (PNC) and between PNC and grain protein content (GPC) were investigated using multi-temporal moderate-resolution imaging spectroradiometer (MODIS) data at the different stages of winter wheat in Linfen (Shanxi, P. R. China). The anticipating model for GPC of winter wheat was also established by the approach of NDVI at the different stages of winter wheat. The results showed that the spectrum models of PNC passed F test. The NDVI4.14 regression effect of PNC model of irrigated winter wheat was the best, and that in dry land was NDVI4.30. The PNC of irrigated and dry land winter wheat were significantly (P<0.01) and positively correlated to GPC. Both of protein spectral anticipating model of irrigated and dry land winter wheat passed a significance test (P<0.01). Multiple anticipating models (MAM) were established by NDVI from two periods of irrigated and dry land winter wheat and PNC to link GPC anticipating model. The coefficient of determination R2 (R) of MAM was greater than that of the other two single-factor models. The relative root mean square error (RRMSE) and relative error (RE) of MAM were lower than those of the other two single-factor models. Therefore, test effects of multiple proteins anticipating model were better than those of single-factor models. The application of multiple anticipating models for predication of protein content (PC) of irrigated and dry land winter wheat was more accurate and reliable. The regionalization analysis of GPC was performed using inverse distance weighted function of GIS, which is likely to provide the scientific basis for the reasonable winter wheat planting in Linfen city, China.

Abstract. Earth observation via remote sensing imagery provides a fast way to define alteration levels. In this work 12 stands of Araucaria-Nothofagus forests were selected in southern Chile, which represented four alteration levels: (i) None (ii) Low (iii) Medium and (iv) High. The stands were surveyed measuring 379 field plots and Google Earth Engine was used to collect a composite of Sentinel-2 images over a one-year range, from June 2019 to June 2020. The following approaches were tested: (i) aggregating the normalized difference vegetation index (NDVI) of each image and selecting the 95th and 99th percentile values of NDVI for each pixel; (ii) creating a composite imagery with best pixels over one year timeline using NDVI as weighting factor and NDVI value band itself (NDVI) – this is similar to the 99th percentile in the previous point, but with maximum values of NDVI; (iii) aggregating the composite as in the previous approach, but using the full spectral information of Sentinel-2 and then random forest machine learning for classification over alteration areas with k-fold validation with k=5. Results show that the 95th and 99th percentile of NDVI values from approach (i) do not discriminate the four classes correctly. The maximum NDVI from approach can distinguish all four classes. It must be noted through that statistical significance does not necessarily imply a strong practical significance; medium and high alterations have very similar NDVI distributions. Random forest results provided an F-score for each class higher than 80% except for the “medium alteration” class.

A 16-year time-series (1982-1997) of monthly maximum Normalized Difference Vegetation Index (NDVI) values derived from coarse-scale Advanced Very High Resolution Radiometer (AVHRR) satellite images was investigated for its application in identifying agricultural regions sensitive to El Ni@o-Southern Oscillation (ENSO) impacts. Cumulative NDVI values over the growing season were used to test the temporal and spatial sensitivity of rainfed agricultural regions in the US cornbelt to ENSO climate teleconnections. The correlation (R) between NDVI anomalies and yield anomalies aggregated to the Agricultural Statistics Division (ASD) level was 0.59 (significant at the 99% level). Sea surface temperatures (SSTs) from the NINO3 region (between +5.0° and m 5.0° latitude and 90.0° and 180.0° west longitude in the Pacific Ocean), an indicator of the ENSO phenomenon, were significantly but weakly correlated with growing season NDVI anomalies, precipitation anomalies, and yield anomalies. Two classification schemes for El Niño, La Niña and Neutral years related to the crop growing season were applied--one based on simultaneous Pacific NINO3 SSTs (spring-summer), and one based on following winter Pacific NINO3 SSTs. The strongest differences in Vegetation Condition Index (VCI) spatial patterns among the three ENSO categories were found using the following winter SST classification. Classification of ENSO years is a key issue in analysing ENSO impacts on agriculture as represented by the NDVI, because the regularity of annual agricultural seasons is not synchronous with the quasi-regular Pacific SST cycles.

BACKGROUND AND AIM: Exposure to green space (natural vegetation) has been linked to improved reproductive outcomes, potentially through behavioral, psychological, and physiological mechanisms. However, there has been no study of the role of green space in gynecologic health. We examined the association between residential green space and menstrual cycle characteristics in a cohort of reproductive-aged women. METHODS: We performed a cross-sectional analysis within Pregnancy Study Online (PRESTO), a web-based prospective cohort study of women aged 21-45 years who reside in the United States or Canada. We included 7,733 women who enrolled during June 2013-April 2019. We geocoded their residential addresses and calculated annual maximum normalized difference vegetation index (NDVI) at 30 meter resolution within 250 meters around their residences to quantify green space exposure. Women reported on menstrual cycle regularity, cycle length, bleed length, heaviness of bleed, and intensity of menstrual pain. We used log-binomial regression models to estimate prevalence ratios (PRs) and 95% confidence intervals (CIs), adjusting for sociodemographic, lifestyle, and neighborhood characteristics, including population density and census tract median household income. RESULTS:The median (interquartile range) NDVI within 250 meters was 0.64 (0.52, 0.74). Low residential green space was associated with a higher prevalence of long bleeds (≥6 days) and severe period pain (medication and bed rest required). The PRs comparing the lowest levels of green space (NDVI 0.2) with the highest levels of green space (NDVI ≥0.8) were 1.43 (95% CI: 0.99, 2.08) for long bleeds and 1.71 (95% CI: 1.01, 2.89) for severe period pain. However, there was little evidence of dose-response associations. Other menstrual cycle characteristics were not appreciably associated with NDVI. CONCLUSIONS:Women with the lowest residential green space were more likely to have long menstrual bleeds and severe menstrual pain, indicating a potential role for green space in menstrual health. KEYWORDS: reproductive outcomes, green space, built environment, epidemiology

Leaf Area Index (LAI) is an important biophysical parameter necessary to infer vegetation vigour, seasonal vegetation variability and different physiological and biochemical processes of vegetation. Gap fraction analysis has been carried out to estimate plot‐wise LAI. Tropical dry deciduous forests of the study area can be categorized into two prominent phases—growing and senescent phases. Efforts have been made to observe the relationship between ground based LAI values and satellite derived parameters, such as radiance values and also different vegetation indices, namely the Normalized Difference Vegetation Index (NDVI), maximum, minimum, amplitude, sum and radiance NDVI values for both the phases. Multi‐temporal IRS 1C WiFS data have been used. WiFS provides information in two bands: red (0.62–0.68 µm) and near‐infrared (0.77–0.86 µm). All the images from the representative months have been corrected radiometrically and geometrically. NDVI values have been derived for all the representative months. Among various parameters maximum NDVI values showed better relationships with LAI for both the phases. For the growing phase R 2 (coefficient of determination) was 0.79, whereas for the senescent phase it was 0.48. These empirical relationships have been used to estimate LAI at a regional level. The LAI estimate for growing and senescent phases ranged from <1 to 4.25.

Crop type identification plays an important role in extracting crop acreage, assessing crop growth and arable land productivity. In this study, the main crops (winter wheat, summer maize and cotton) of Shandong Province as research objects, and the SPOT_VGT normalized difference vegetation index (NDVI) remote sensing datasets from 1999 to 2011 covering Shandong Province were acquired. The NDVI characteristic curves of typical features were extracted by combining the SPOT_VGT NDVI time series datasets, the HJ-1B image and the phenological information. Moreover, the reasonable dynamic thresholds were settled, the non-cultivated land areas were removed and the crop patterns and the crop types were identified based on the annual NDVI variation and the phenological information of the typical features. The accuracy assessment was performed through the spatial contrast and quantitative description. The overall accuracy is 77.10% in the spatial accuracy assessment compared with standard land cover classification ...

The Google Earth Engine (GEE) has emerged as an essential cloud-based platform for land-cover classification as it provides massive amounts of multi-source satellite data and high-performance computation service. This paper proposed an automatic land-cover classification method using time-series Landsat data on the GEE cloud-based platform. The Moderate Resolution Imaging Spectroradiometer (MODIS) land-cover products (MCD12Q1.006) with the International Geosphere–Biosphere Program (IGBP) classification scheme were used to provide accurate training samples using the rules of pixel filtering and spectral filtering, which resulted in an overall accuracy (OA) of 99.2%. Two types of spectral–temporal features (percentile composited features and median composited monthly features) generated from all available Landsat Thematic Mapper (TM) and Enhanced Thematic Mapper Plus (ETM+) data from the year 2010 ± 1 were used as input features to a Random Forest (RF) classifier for land-cover classification. The results showed that the monthly features outperformed the percentile features, giving an average OA of 80% against 77%. In addition, the monthly features composited using the median outperformed those composited using the maximum Normalized Difference Vegetation Index (NDVI) with an average OA of 80% against 78%. Therefore, the proposed method is able to generate accurate land-cover mapping automatically based on the GEE cloud-based platform, which is promising for regional and global land-cover mapping.

Extraction of Agricultural Phenological Parameters of Sri Lanka Using MODIS, NDVI Time Series Data