Official Agricultural Statistics Research Articles

Using Auxiliary Information to Improve Agricultural Statistics – Advantages of the Calibration Approach over Poststratification Weights

Official statistics are often based on samples representing a certain population. Because participation in a sample is usually voluntary, bias might result from so-called non-sampling errors such as nonresponse. Weighting procedures are intended to correct these errors by assigning a certain weight to each observation in the sample. In many official agricultural statistics, such as the Bavarian Agricultural Report, poststratification is used. In this process, the population is stratified according to different dimensions (e.g. farm type, farm location and farm size) and weights are assigned to all farms in a stratum so that the sum of the weights in that stratum corresponds to the number of observations in that stratum in the population. However, when estimating the population average, important characteristics (such as the farm size) may still be biased. Using a Bavarian farm sample, the present study shows how the so-called calibration approach, utilising auxiliary variables to adjust weights, outperforms the poststratification procedure in terms of estimating important population characteristics.

The Estimation of the Long-Term Agricultural Output with a Robust Machine Learning Prediction Model

Recently, annual agricultural data have been highly volatile as a result of climate change and national economic trends. Therefore, such data might not be enough to develop good agricultural policies for stabilizing agricultural output. A good agricultural output prediction model to assist agricultural policymaking has thus become essential. However, the highly volatile data would affect the prediction model’s performance. For this reason, this study proposes a marriage in honey bees optimization/support vector regression (MBO/SVR) model to minimize the effects of highly volatile data (outliers) and enhance prediction accuracy. We verified the performance of the MBO/SVR model by using the annual total agricultural output collected from the official Agricultural Statistics Yearbook of the Council of Agriculture, Taiwan. Taiwan’s annual total agricultural output integrates agricultural, livestock and poultry, fishery, and forest products. The results indicated that the MBO/SVR model had a lower mean absolute percentage error (MAPE), root mean square percentage error (RMSPE), and relative root mean squared error (r-RMSE) than those of the models it was compared to. Furthermore, the MBO/SVR model predicted long-term agricultural output more accurately and achieved higher directional symmetry (DS) than the other models. Accordingly, the MBO/SVR model is a robust, high-prediction-accuracy model for predicting long-term agricultural output to assist agricultural policymaking.

New Players on a Tough Field

Abstract The pivotal aim of this research is to identify new entrants to mountain farming, their routes into practicing and resulting environmental impacts. Following an actor-network approach, this multi-regional case study was conducted in two remote regions of the Western Austrian Alps and in a bordering Alpine region in South Tyrol (Italy). All data was generated using semi-structured interviews on site and official agricultural statistics. Both autochthons and foreign newcomers to mountain farming display a great deal of idealism to fulfil their agricultural dreams. Their self-determination and bid for independence from the agricultural industry affects their thinking and decisions on housing, farming and collaborations, resulting in far-reaching social and spatial impacts on the sociocultural landscape. In implementing sustainable and extensive farming practices, new entrants act as role models and thus highlight alternatives beyond the prevailing mountain agricultural regime.

LATVIAN SEMI-NATURAL GRASSLANDS IN THE SOCIO-ECOLOGICAL EXTINCTION VORTEX: THE EVIDENCE AND RESEARCH GAPS

INTRODUCTION
 Semi-natural grasslands depend on active management by certain societal actors, and can therefore be defined as socio-ecological systems (Herzon et al., 2021). Semi-natural grasslands of the European Boreal biogeographical region are considered to be caught in a socio-ecological extinction vortex (Herzon et al., 2022). Our aim was to look at the evidence for the extinction vortex of Latvian semi-natural grasslands at the national level, and to describe the main directions in which research is insufficient.
 MATERIAL AND METHODS
 We conducted a scientific literature review and collected official agricultural and habitat conservation statistics.
 RESULTS
 Published studies are dominated by ecological aspects of semi-natural grasslands in Latvia. At the national level, there is little evidence available about all four interlinked and mutually reinforcing socio-economic processes operating in the extinction vortex (Herzon et al., 2022): 1) agricultural production; 2) policy, research and development; 3) vocational education in the fields of agricultural sciences and 4) the public’s experience with semi-natural habitats.
 DISCUSSION
 The success of semi-natural grassland conservation depends on how well social-ecological system (SES) thinking has been applied through conservation planning and implementation (Krievins et al. 2018; Fischer et al. 2021). Therefore, conservation should approach semi-natural grasslands as socio-ecological systems rather than merely ecosystems. In Latvia, this approach has not yet been fully appreciated.
 CONCLUSIONS
 There is a lack of understanding of semi-natural grasslands as socio-ecological systems in Latvia both in the governmental and non-governmental sector, as well as in the research community. The main research gap is in inter- and transdisciplinary approaches to semi-natural grassland conservation.

Does the Granary County Subsidy Program Lead to manipulation of grain production data in China?

Manipulation of food production data could lead to catastrophic social and economic consequences. The accuracy of official agricultural statistics has long been questioned in China. This paper studies the linkage between agricultural production data manipulation and the Granary County Subsidy Program (GCSP). Since 2005, Chinese government gave subsidies to those counties with five-year average grain production between 1998 and 2002 more than 200 thousand tons to encourage these local governments to give priority to grain production. The prospective counties with grain production slightly below the threshold may have incentives to over-report their grain production. Based on the McCrary (2008)'s density test, this paper provides suggestive evidence of over-reporting grain production caused by the GCSP in 2005, 2006 and 2008, though the over-reporting rates are only 3%, 2%, and 1.7% respectively. The policy implication would be that fiscal distribution rules of a central government should avoid data manipulation incentives in local governments, particularly should cut the linkage to the data which are self-reported by the local governments.

On-farm compliance costs with the EU-Nitrates Directive: A modelling approach for specialized livestock production in northwest Germany

In the EU, several environmental regulations aim at protecting the environment from agricultural nitrogen and phosphorus losses. The German regulation on farmers' nutrient management, especially implementing the EU Nitrates Directive, was revised in 2017. It comprises considerable tightening of numerous measures and costs for farmers to comply with. We provide the first systematic farm-level analysis of compliance costs of the recent revision in a case study for the federal state of North Rhine-Westphalia. To do so, we apply a bio-economic optimisation model at farm level to a representative sample of specialized dairy and pig farms. The sample is derived by Latin Hypercube sampling based on the observed distribution of farm characteristics from official agricultural statistics. Modelling results are evaluated by grouping of farms and a statistical meta-model. Results show highly heterogeneous compliance costs reaching from 0 to 2.66 Euro per pig and 0 to 0.83 cent per kg milk. 47.3% of pig and 38.4% of dairy farms do not face any costs. Pig farms with high compliance costs are characterized by high stocking density, the absence of low-emission manure application techniques and phosphorus-enriched soils. Dairy farms with high compliance costs have no low-emission manure application techniques and a high share of grassland. For dairy farms, stricter thresholds for nutrient application do not cause any compliance costs. The meta-model reveals the large effect of prices and assumptions regarding the fertilizer management on compliance costs. Results are of relevance beyond the case study area as other regions in the EU have a similar agricultural structure and need to fulfill the same EU directives. Policymakers need to be aware that high compliance costs increase the incentive of non-compliance and also consider heterogeneous impacts when designing complementary policies. Future research should focus on long-term adaption of farmers and include transaction costs as well as technical progress.

Increasing concentration of major crops in China from 1980 to 2011

ABSTRACTThe concentration of crop cultivation can be measured in terms of spatial clustering and of inequality in the distribution of the cropland area. We used official agricultural statistics at the county level (N = 2,354) for each year from 1980 to 2011 for all of China to analyse the changes in spatial clustering and inequality of overall cropland and of the harvested areas of the five major crops (rice, maize, wheat, soybean, and potato). We quantified the spatial clustering with global and local Moran’s I and assessed the inequality in the distribution of crop cultivation with the generalized entropy index. The results showed that the cropland area and harvested areas of the major crops indeed became more homogeneous over time, and the major crops concentrated in fewer areas and in the major historic breadbaskets. Increasing concentration may offer opportunities in specialization and positive agglomeration effects but can reduce the resilience of food systems and agricultural sustainability.

Canopy temperatures distribution over soybean crop fields using satellite data in the Amazon biome frontier

ABSTRACTDuring the studied time window, between 2003 and 2010, there was an important increase of land use conversion into new soybean areas (first-time-use) in Mato Grosso state (MT) in Brazil. Uncertainties of future scenario of Brazilian agriculture and increase in the frequency of extreme events, such as the occurrence of high temperatures, is highly likely to produce yield loss on summer crops. The MT is the largest producer of soybeans and accounted for 28.2% of the national production in 2013. The objective of this study was to investigated specific characterization of land surface temperature distribution over the soybean crop fields canopies (canopy-LST) due to massive land use conversion into new soybean areas and its impacts on yield. Satellite imagery data from Aqua and Terra/MODIS sensors (Moderate Resolution Imaging Spectroradiometer) were compared to official agricultural statistics covering eight densely cultivated regions in the studied period. Results show that within the period from flowering to grain filling canopy-LST exhibits a non-negligible relation to yield. It is expected an additional loss of 4.9% on soybean yield for each 1oC of canopy-LST above the obtained optimal level of canopy-LST with 28.4oC, associated to the higher yield averages. The difference between overall average of canopy-LST and air temperature was found 4.2 oC.

MasAgro o MIAF ¿Cuál es la opción para modernizar sustentablemente la agricultura tradicional de México


 
 
 Las estadísticas agropecuarias oficiales de los últimos 35 años (SIAP, 2017) sugieren que el avance en el objetivo de MasAgro de incrementar la producción de maíz de temporal y su rendimiento a escala nacional es inapreciable. Los autores de este ensayo calificamos como poco probable que ese objetivo sea alcanzado en los cinco años que restan del programa. Se discute en este ensayo al sistema tecnológico milpa intercalada en árboles frutales (MIAF), como alternativa a MasAgro, para hacer más productiva y sustentable la agricultura tradicional en pequeño de México. El MIAF ha sido desarrollado por el Instituto Nacional de Investigaciones Forestales, Agrícolas y Pecuarias y el Colegio de Postgraduados en Ciencias Agrícolas en más de 30 años de colaboración. Su diseño persigue la intensificación del paradigma de la agricultura tradicional (PAT) en pequeño. Retiene ventajas clave de la milpa histórica como a) resiliencia apoyada en la biodiversidad; y b) eficiencia relativa de la tierra superior a la unidad. También aprovecha la tecnología de la terraza de muro vivo formada con árboles frutales, para la protección del suelo contra la erosión y para acrecentar el ingreso y empleo familiares. Se concluye que MIAF es la tecnología indicada para modernizar la agricultura tradicional en pequeño de México.
 
 

A Remote Sensing Approach for Regional-Scale Mapping of Agricultural Land-Use Systems Based on NDVI Time Series

In response to the need for generic remote sensing tools to support large-scale agricultural monitoring, we present a new approach for regional-scale mapping of agricultural land-use systems (ALUS) based on object-based Normalized Difference Vegetation Index (NDVI) time series analysis. The approach consists of two main steps. First, to obtain relatively homogeneous land units in terms of phenological patterns, a principal component analysis (PCA) is applied to an annual MODIS NDVI time series, and an automatic segmentation is performed on the resulting high-order principal component images. Second, the resulting land units are classified into the crop agriculture domain or the livestock domain based on their land-cover characteristics. The crop agriculture domain land units are further classified into different cropping systems based on the correspondence of their NDVI temporal profiles with the phenological patterns associated with the cropping systems of the study area. A map of the main ALUS of the Brazilian state of Tocantins was produced for the 2013–2014 growing season with the new approach, and a significant coherence was observed between the spatial distribution of the cropping systems in the final ALUS map and in a reference map extracted from the official agricultural statistics of the Brazilian Institute of Geography and Statistics (IBGE). This study shows the potential of remote sensing techniques to provide valuable baseline spatial information for supporting agricultural monitoring and for large-scale land-use systems analysis.

Long-term yield trends of insect-pollinated crops vary regionally and are linked to neonicotinoid use, landscape complexity, and availability of pollinators

Time series data on crop yields for two main wind-pollinated crops (barley and wheat) and for three crops benefitting from insect pollination (turnip rapeseed, caraway, and black currant), were compiled from official agricultural statistics. In Finland, these statistics are available at aggregate national level, and at the level of each of the 15 provinces of the country. Yields of wind-pollinated crops have steadily increased in Finland, while yields of insect-pollinated crops have been highly variable. The largest crop benefitting from insect pollination is turnip rapeseed, which shows first a clear tendency to increased yields from 1980 to 1993, after which there has been a continuous decline in yields at the national average level. Regionally, the trends in turnip rapeseed yield show large variation, so that in six provinces of Finland, the trend has been significantly decreasing; in five provinces, there has been no significant trend; and in two provinces, there has been a significant linear increase in yields. Yield trends in the two other insect-pollinated crops, caraway and black currants, show similar trend variations. However, at the national average level, caraway yields show no significant trend, while black currant yields have increased during the past 6 years. The possible impact on the trends of insect-pollinated crops of three explanatory variables was analyzed. Significant linear correlation was found between the yield trends (slope of the trends) in rapeseed, and the extent of using neonicotinoid seed dressing in the provinces; the magnitude of yield decline in turnip rapeseed increased, as the use of neonicotinoid seed dressing increased. Similar significant linear correlation was found for the magnitude of yield decline in turnip rapeseed and the complexity of the agricultural landscape in each province; yield trend changed from negative to positive as the proportion of agricultural land of the total terrestrial land area declined from 28% to below 10%. The availability of honey bee colonies with respect to the growing area of crops benefitting from insect pollination also had a linear, significant impact on turnip rapeseed yield trends: yields tended to decline in provinces, where the supply of managed pollinators with respect to demand was low, but tended to increase in provinces, where the number of honey bee colonies were over 30% of the estimated demand. As neither the landscape complexity (proportion of arable land of total terrestrial land area), nor the number of honey bee colonies for pollination have changed significantly over the past 10–20 years, these factors cannot explain the observed differences in the yield trends of the examined insect-pollinated crops. It appears that only the uptake of neonicotinoid insecticide seed dressing about 15 years ago can explain the crop yield declines in several provinces, and at the national level for turnip rapeseed, most likely via disruption of pollination services by wild pollinators.

Model for soybean production forecast based on prevailing physical conditions

Abstract: The objective of this work was to evaluate the reliability of the physiological meaning of the enhanced vegetation index (EVI) data for the development of a remote sensing-based procedure to estimate soybean production prior to crop harvest. Time-series data from the moderate resolution imaging spectroradiometer (Modis) were applied to investigate the relationship between local yield fluctuations of soybean and the prevailing physically-driven conditions in the state of Mato Grosso, located in the south of the Brazilian Amazon. The developed methodology was based on the coupled model (CM). The CM provides production estimates for early January, using images from the maximum crop development period. Production estimates were validated at three different spatial scales: state, municipality, and local. At the state and municipality levels, the results obtained from the CM were compared with official agricultural statistics from Instituto Brasileiro de Geografia e Estatística and Companhia Nacional de Abastecimento, from 2001 to 2011. The coefficients of determination ranged from 0.91 to 0.98, with overall result of R2=0.96 (p≤0.01), indicating that the model adheres to official statistics. At the local level, spatially distributed data were compared with production data from 422 crop fields. The coefficient of determination (R2=0.87) confirmed the reliability of the EVI for its applicability on remote sensing-based models for soybean production forecast.

Crop Monitoring Using Vegetation and Thermal Indices for Yield Estimates: Case Study of a Rainfed Cereal in Semi-Arid West Africa

For the semiarid Sahelian region, climate variability is one of the most important risks of food insecurity. Field experimentations as well as crop modeling are helpful tools for the monitoring and the understanding of yields at local scale. However, extrapolation of these methods at a regional scale remains a demanding task. Remote sensing observations appear as a good alternative or addition to existing crop monitoring systems. In this study, a new approach based on the combination of vegetation and thermal indices for rainfed cereal yield assessment in the Sahelian region was investigated. Empirical statistical models were developed between MODIS NDVI and LST variables and the crop model SARRA-H simulated aboveground biomass and harvest index in order to assess each component of the yield equation. The resulting model was successfully applied at the Niamey Square Degree (NSD) site scale with yield estimations close to the official agricultural statistics of Niger for a period of 11 years (2000–2011) ( $r = 0.82,\;\text{p-value} ). The combined NDVI and LST indices-based model was found to clearly outperform the model based on NDVI alone ( $r = 0.59,\;\text{p-value} ). In areas where access to ground measurements is difficult, a simple, robust, and timely satellite-based model combining vegetation and thermal indices from MODIS and calibrated using crop model outputs can be pertinent. In particular, such a model can provide an assessment of the year-to-year yield variability shortly after harvest for regions with agronomic and climate characteristics close to those of the NSD study area.

Where have all the pigs gone? Inconsistencies in pork statistics in China

Official agricultural statistics for China are subject to major inconsistencies and have long been questioned by researchers. The major problem with Chinese meat statistics is that reported meat supply is far greater than consumption, and this is particularly true for pork. Factors contributing to the gap between pork supply and consumption include production over-reporting, loss and waste in the pork supply chain, pork consumed away from home (FAFH), and a mismatch in the Chinese rural household survey between food and mouths (i.e. migrant workers and boarding students who are counted as rural household members but live in urban areas for much of the year). Our estimates indicate that over-reporting of pork production has declined substantially since 2003, but it is still significant and is the largest contributor to the gap between reported supply and consumption. Our estimates also indicate that pork consumption is significantly under-estimated because of FAFH and the rural mismatch between food and mouths. Reforms to the agricultural statistical system should be considered that increase the incentives to report accurate production statistics. Statistics are currently based on reports from local officials who have incentives to inflate production figures so as to improve their performance reviews and prospects for promotion, or they are overseen by local statistical personnel appointed by local governments.

Assessing the MODIS crop detection algorithm for soybean crop area mapping and expansion in the Mato Grosso state, Brazil.

Estimations of crop area were made based on the temporal profiles of the Enhanced Vegetation Index (EVI) obtained from moderate resolution imaging spectroradiometer (MODIS) images. Evaluation of the ability of the MODIS crop detection algorithm (MCDA) to estimate soybean crop areas was performed for fields in the Mato Grosso state, Brazil. Using the MCDA approach, soybean crop area estimations can be provided for December (first forecast) using images from the sowing period and for February (second forecast) using images from the sowing period and the maximum crop development period. The area estimates were compared to official agricultural statistics from the Brazilian Institute of Geography and Statistics (IBGE) and from the National Company of Food Supply (CONAB) at different crop levels from 2000/2001 to 2010/2011. At the municipality level, the estimates were highly correlated, with R 2 = 0.97 and RMSD = 13,142 ha. The MCDA was validated using field campaign data from the 2006/2007 crop year. The overall map accuracy was 88.25%, and the Kappa Index of Agreement was 0.765. By using pre-defined parameters, MCDA is able to provide the evolution of annual soybean maps, forecast of soybean cropping areas, and the crop area expansion in the Mato Grosso state.

Mapping the Spatial Distribution of Winter Crops at Sub-Pixel Level Using AVHRR NDVI Time Series and Neural Nets

For large areas, it is difficult to assess the spatial distribution and inter-annual variation of crop acreages through field surveys. Such information, however, is of great value for governments, land managers, planning authorities, commodity traders and environmental scientists. Time series of coarse resolution imagery offer the advantage of global coverage at low costs, and are therefore suitable for large-scale crop type mapping. Due to their coarse spatial resolution, however, the problem of mixed pixels has to be addressed. Traditional hard classification approaches cannot be applied because of sub-pixel heterogeneity. We evaluate neural networks as a modeling tool for sub-pixel crop acreage estimation. The proposed methodology is based on the assumption that different cover type proportions within coarse pixels prompt changes in time profiles of remotely sensed vegetation indices like the Normalized Difference Vegetation Index (NDVI). Neural networks can learn the relation between temporal NDVI signatures and the sought crop acreage information. This learning step permits a non-linear unmixing of the temporal information provided by coarse resolution satellite sensors. For assessing the feasibility and accuracy of the approach, a study region in central Italy (Tuscany) was selected. The task consisted of mapping the spatial distribution of winter crops abundances within 1 km AVHRR pixels between 1988 and 2001. Reference crop acreage information for network training and validation was derived from high resolution Thematic Mapper/Enhanced Thematic Mapper (TM/ETM+) images and official agricultural statistics. Encouraging results were obtained demonstrating the potential of the proposed approach. For example, the spatial distribution of winter crop acreage at sub-pixel level was mapped with a cross-validated coefficient of determination of 0.8 with respect to the reference information from high resolution imagery. For the eight years for which reference information was available, the root mean squared error (RMSE) of winter crop acreage at sub-pixel level was 10%. When combined with current and future sensors, such as MODIS and Sentinel-3, the unmixing of AVHRR data can help in the building of an extended time series of crop distributions and cropping patterns dating back to the 80s.

Spectral Model for Soybean Yield Estimate Using MODIS/EVI Data

Attaining reliable and timely agricultural estimates is very important everywhere, and in Brazil, due to its characteristics, this is especially true. In this study, estimations of crop production were made based on the temporal profiles of the Enhanced Vegetation Index (EVI) obtained from Moderate Resolution Imaging Spectroradiometer (MODIS) images. The objective was to evaluate the coupled model (CM) performance of crop area and crop yield estimates based solely on MODIS/EVI as input data in Rio Grande do Sul State, which is characterized by high variability in seasonal soybean yields, due to different crop development conditions. The resulting production estimates from CM were compared to official agricultural statistics of Brazilian Institute of Geography and Statistics (IBGE) and the National Company of Food Supply (CONAB) at different levels from 2000/2001 to 2010/2011 crop years. Results obtained with CM indicate that its application is able to generate timely production estimates for soybean both at municipality and local levels. Validation estimates with CM at State level obtained R2 = 0.95. Combining all cropping years at municipality level, estimates were highly correlated to official statistics from IBGE, with R2 = 0.91 and RMSD = 10,840 tons. Spatially interpolated comparisons of yield maps obtained from the CM estimates and IBGE data also showed visual similarity in their spatial distribution. Local level comparisons were performed and presented R2 = 0.95. Implications of this work point out that time-series analysis of production estimates are able to provide anticipated spatial information prior to the soybean harvest.

SOYBEAN CROP AREA ESTIMATION AND MAPPING IN MATO GROSSO STATE, BRAZIL

Abstract. Evaluation of the MODIS Crop Detection Algorithm (MCDA) procedure for estimating historical planted soybean crop areas was done on fields in Mato Grosso State, Brazil. MCDA is based on temporal profiles of EVI (Enhanced Vegetation Index) derived from satellite data of the MODIS (Moderate Resolution Imaging Spectroradiometer) imager, and was previously developed for soybean area estimation in Rio Grande do Sul State, Brazil. According to the MCDA approach, in Mato Grosso soybean area estimates can be provided in December (1st forecast), using images from the sowing period, and in February (2nd forecast), using images from sowing and maximum crop development period. The results obtained by the MCDA were compared with Brazilian Institute of Geography and Statistics (IBGE) official estimates of soybean area at municipal level. Coefficients of determination were between 0.93 and 0.98, indicating a good agreement, and also the suitability of MCDA to estimations performed in Mato Grosso State. On average, the MCDA results explained 96% of the variation of the data estimated by the IBGE. In this way, MCDA calibration was able to provide annual thematic soybean maps, forecasting the planted area in the State, with results which are comparable to the official agricultural statistics.

Soybean crop area estimation by Modis/Evi data

The objective of this work was to develop a procedure to estimate soybean crop areas in Rio Grande do Sul state, Brazil. Estimations were made based on the temporal profiles of the enhanced vegetation index (Evi) calculated from moderate resolution imaging spectroradiometer (Modis) images. The methodology developed for soybean classification was named Modis crop detection algorithm (MCDA). The MCDA provides soybean area estimates in December (first forecast), using images from the sowing period, and March (second forecast), using images from the sowing and maximum crop development periods. The results obtained by the MCDA were compared with the official estimates on soybean area of the Instituto Brasileiro de Geografia e Estatística. The coefficients of determination ranged from 0.91 to 0.95, indicating good agreement between the estimates. For the 2000/2001 crop year, the MCDA soybean crop map was evaluated using a soybean crop map derived from Landsat images, and the overall map accuracy was approximately 82%, with similar commission and omission errors. The MCDA was able to estimate soybean crop areas in Rio Grande do Sul State and to generate an annual thematic map with the geographic position of the soybean fields. The soybean crop area estimates by the MCDA are in good agreement with the official agricultural statistics.

Forecasting crop production using satellite-based vegetation health indices in Kansas, USA

This article shows the results of early crop yield prediction from remote-sensing data. The study was carried out in Kansas, USA. The methodology proposed allows the estimation of winter wheat (WW), sorghum and corn yields 3–4 months before harvest. The procedure uses the vegetation health (VH) indices (vegetation condition index (VCI) and temperature condition index (TCI)) computed for each pixel and week over a 21-year period (1985–2005) from the Advanced Very High Resolution Radiometer (AVHRR) data. Over this period, a strong correlation was found between crop yield and VH indices during the weather-related critical period of crop development, which controls much final crop productivity. The 3-month advanced yield forecasts were independently compared with official agricultural statistics, showing that the estimation errors for WW, sorghum and corn were 8%, 6% and 3%, respectively. Implementing the 3–4 months lead forecast in operational practice will aid farmers to mitigate weather vagaries using irrigation, diseases/insects control, application of fertilizers and so on during a growing season and will help decision-makers to regulate marketing strategies, import/export and price policies and address food security issues.