Crop condition and yield simulations using Landsat and MODIS

Monitoring crop condition and yields at regional scales using imagery from operational satellites remains a challenge because of the problem in scaling local yield simulations to the regional scales. NOAA AVHRR satellite imagery has been traditionally used to monitor vegetation changes that are used indirectly to assess crop condition and yields. Additionally, the 1-km spatial resolution of NOAA AVHRR is not adequate for monitoring crops at the field level. Imagery from the new MODIS sensor onboard the NASA Terra satellite offers an excellent opportunity for daily coverage at 250-m resolution, which is adequate to monitor field sizes are larger than 25 ha. A field study was conducted in the predominantly corn and soybean area of Iowa to evaluate the applicability of the 8-day MODIS composite imagery in operational assessment of crop condition and yields. Ground-based canopy reflectance and leaf area index (LAI) measurements were used to calibrate the models. The MODIS data was used in a radiative transfer model to estimate LAI through the season. LAI was integrated into a climate-based crop simulation model to scale from local simulation of crop development and responses to a regional scale. Simulations of corn and soybean yields at a 1.6×1.6-km² grid scale were comparable to county yields reported by the USDA-National Agricultural Statistics Service (NASS). Weekly changes in soil moisture for the top 1-m profile were also simulated as part of the crop model as one of the critical parameters influencing crop condition and yields.     

NOAA-AVHRR NDVI decomposition and subpixel classification using linear mixing in the Argentinean Pampa

Crop Normalized Difference Vegetation Index (NDVI) time profiles and crop acreage estimates were derived from the application of linear mixture modelling to Advanced Very High Resolution Radiometer (AVHRR) data over a test area in the southern part of the Pampa region, Argentina. Bands 1 and 2 from seven AVHRR scenes (June to January 1991) were combined to produce fraction images of winter crops, summer crops and pastures. A Landsat Thematic Mapper (TM) scene of the region was classified and superimposed to the AVHRR Local Area Coverage (LAC) data by means of a correlation technique. Each class signature was extracted by regressing the AVHRR response on the cover types proportions, estimated from Landsat-TM data, over sets of calibration windows. The crop NDVI profiles were hence derived from the class signatures in bands 1 and 2. These profiles appeared consistent with the cover types, but variability depending on the set of windows was noted. The assessment of the class signatures was indirectly accomplished through the subpixel classifications of the AVHRR data, performed using the different sets of class spectra. Although some discrepancies between AVHRR and Landsat–TM estimates were observed at the individual window level, the classification results compared quite well on a regional scale with Landsat–TM estimates: crop acreage was estimated to an overall accuracy ranging from 89 to 95 per cent according to the spectra used in the classification. Definitely, the proposed methodology should permit a better exploitation of the temporal resolution of AVHRR data in both the areas of yield prediction and vegetation classification. Furthermore, the perational application of such a methodology for crop monitoring will undoubtedlybe facilitated with the coming sensor systems such as the ModerateResolution Imaging Spectroradiometer (MODIS), the SPOT Vegetation Monitoring Instrument or the ‘Satelite Argentino Cientifico’ (SAC–C).     

Scale considerations in vegetation monitoring using AVHRR data

Abstract

The Advanced Very High Resolution Radiometer (AVHRR) is currently the only operational remote sensing system capable of providing global daily data which can be used for vegetation monitoring. These data are available with resolution cell sizes ranging from around one to 20 km on a side, though the temporal and spatial extent of cover at each resolution is variable. In this paper Normalized Difference Vegetation Index temporal curves derived from AVHRR at different resolutions are compared over both agricultural and natural tropical vegetation types. For the agricultural regions the length of growing season and major breaks of slope associated with key crop development events are equally well shown at coarse and fine resolution. Detailed examination of the curves reveals differences thought to result from temporal changes in landscape structure. Temporal curves derived from AVHRR data at dilTerent spatial resolutions shows that the spatial organization of both agricultural and natural landscapes, tropical forest in this case, changes throughout a single season. Transitions across major ecological zones are detected across a range of resolutions, though the undersampling employed in the generation of the coarser resolution products is found to exert some limitations on the spatial representivity of these data; this varies both with geographical area and time. These observations highlight the importance of a consideration of scale when using AVHRR data for vegetation monitoring, and emphasize the need for dilTerent scales of observation (both in temporal and spatial terms) for different problems and at different times of the year.     

Estimation of methane emission from West Siberian wetland by scaling technique between NOAA AVHRR and SPOT HRV

Wetlands are one of the most important ecosystems in the world and at the same time they are presumed to be a source of methane gas, which is one of the most important greenhouse gases. The West Siberian wetlands is the largest in the world and remote sensing techniques can play an important role for monitoring the wetland.High spatial resolution satellite data are effective for monitoring land cover type changes, but can't cover a wide area because of a narrow swath width. On the other hand, global scale data are indispensable in covering a large area, but are too coarse to get the detailed information due to the low spatial resolution. It is necessary to devise a method for the fusion of the data with different spatial resolutions for monitoring the scale-differed phenomena.In this paper, firstly, a SPOT HRV image near Plotnikovo mire was used to map four wetland ecosystems (birch forest, conifer forest, forested bog and open bog) supplemented by field observation. Then, spectral mixture analysis was performed between NOAA AVHRR and SPOT HRV data acquired on the same day.Secondly, field observations were scaled up with these different spatial resolution satellite data. Each of the wetland ecosystem coverage fraction at the sub-pixel level was provided by spectral mixture analysis. Field observation shows that the mean rate of CH₄ emission from forested bog and open bog averaged 21.1 and 233.1 (mg CH₄/m²/day), respectively. The methane emission from the area was estimated by multiplying these average methane emission rates and the fraction coverage in each AVHRR pixel.Finally, the total methane emission over AVHRR coverage was estimated to be 9.46 (10⁹ g CH₄/day) and the mean methane emission over AVHRR coverage was calculated as 59.3 (mg CH₄/m²/day). We could conclude that this mean value is within the probabilistic variability as compared with the airborne measurement results.     

Effet de la reésolution spatiale sur des proprieeacute;teés statistiques des images satellites: une eétude de cas

Abstract

The instruments on such satellites as Landsat or SPOT present the advantage provided by high spatial resolution. This advantage is tempered by their low time resolution. Therefore, it is not always possible to monitor seasonal variations of parameters such as the normalized vegetation index. The AVHRR instrument on board the NOAA satellites has a very high repetitivity but a very low spatial resolution. In our research we proposed to monitor the normalized vegetation index with this low-resolution instrument. It is therefore of interest to examine the relations between high- and low-resolution images for using the AVHRR data as a means of interpolation between two MSS images. This problem is addressed here using satellite images of an important agricultural region in France. In terms of the transformation of the mean radiometric values, it is shown that a linear transformation exists to calculate the AVHRR data from those of the MSS. but the relation presents a strong dependency on the observed scene. The effect on the higher-order statistical properties is studied through the transformation of the images textures by progressively degrading the MSS images. It is shown that a threshold, which depends on the scene, exists on the resolution below which all statistical information disappears.     

Crop specific green area index retrieval from MODIS data at regional scale by controlling pixel-target adequacy

Information on vegetation status can be retrieved from satellite observations by modelling and inverting canopy radiative transfer. Agricultural monitoring and yield forecasting could greatly benefit from such techniques by coupling crop growth models with crop specific information through data assimilation. An indicator which would be particularly interesting to obtain from remote sensing is the total surface of photosynthetically active plant tissue, or green area index (GAI). Currently, the major limitation is that the imagery that can be used operationally and economically over large areas with high temporal frequency has a coarse spatial resolution. This paper demonstrates how it is possible to characterise the regional crop specific GAI range along with its temporal dynamic using MODIS imagery by controlling the degree at which the observation footprints of the coarse pixels fall within the crop-specific mask delineating the target. This control is done by modelling the instrument's point spread function and by filtering out less reliable GAI estimations in both the spatial and temporal dimensions using thresholds on 3 variables: pixel purity, observation coverage and view zenith angle. The difference in performance between MODIS and fine spatial resolution to estimate the median GAI of a given crop over a 40 × 40 km study region can be reduced to a RMSE of 0.053 m²/m². The consistency between fine and coarse spatial resolution GAI estimations suggests a possible instrument synergy whereby the high temporal resolution of MODIS provides the general GAI trajectory and while high spatial resolution can be used to estimate the local GAI spatial heterogeneity.     

Spatial degradation of satellite data

Abstract

Two aspects of spatial degradation of satellite data are examined. The first describes a technique for spatially degrading high-resolution satellite data to produce comparable data sets over a range of coarser resolutions. In this study seven spatial resolution data sets are produced from Landsat Multispectral Scanner (MSS) data resulting in spatial resolutions ranging from 79 m to 4 km applying a spatial filter designed to simulate sensor response. The simulation is demonstrated for part of the Superior National Forest, Minnesota. The second part of the paper examines spatial degradation of coarse resolution data to provide data compression for the production of global-scale data sets. The on-board sampling approach adopted by the National Oceanographic and Atmospheric Administration (NOAA) to produce the Advanced Very High Resolution Radiometer (AVHRR) Global Area Coverage (GAC) data from the 1 km Large Area Coverage (LAC) data, is compared to other sampling procedures. Six sampling procedures were compared for seven terrain types. The GAC sampling procedure provided a relatively poor representation of the 1 km data, particularly for heterogeneous areas. Coefficients of determination for the GAC sampling compared to the original data ranged from 0.49?0.76. Sampling procedures incorporating averaging resulted in a decrease in the variance as compared with the original data. Sampling procedures adopting single-value selection had higher variances and produced data values directly comparable with those from the original data. Sampling scheme design should consider data fidelity requirements as well as the engineering constraints of on-board processing.     

Mean seasonal cycle and evolution of the sea surface temperature from satellite and in situ data in the English Channel for the period 1986–2006

A night-time series of sea surface temperature (SST) of the advanced very high-resolution radiometer (AVHRR) sensors provided by the AVHRR/Pathfinder was analysed over the period 1986–2006 in the English Channel. The studied area is characterized by a strong influence of the bathymetry on the mixing of the water column, mostly through the action of the tide and waves, leading to regional patterns in the SST fields. Another specific aspect of the area is the relatively large number of in situ measurements available from coastal stations. The remotely sensed SST data with fine spatial resolution and high-frequency measurements made at coastal stations have been analysed using a common model. The long-term evolution of SST has been defined in this study through a linear trend while the seasonal evolution has been described through two harmonic functions. The daily satellite SST fields have been estimated over the period 1986–2006 by applying the kriging method to the anomalies calculated from the model. These interpolated temperatures were compared with in situ data, including many coastal stations unreachable at the sensor resolution. To use those coastal stations for comparison and to complement the satellite-derived data set, we defined transfer functions established from fine analysis of the in situ gradients along cross shore transects. The study showed the existence of a long-term warming and that this trend was not homogeneous in the area studied. The central part of the English Channel and the Western part of Brittany show an increase in temperature of about 0.6°C and the Northern part of the Irish and Baltic Sea, included in the studied area, show a maximum increase in the temperature of 1.6°C over the period 1986–2006.     

Assessment of millet yields and production in northern Burkina Faso using integrated NDVI from the AVHRR.

Abstract

It is possible to assess crop yields at the end of the growing season in a semi-arid environment using data from meteorological satellites. This is the result of a work carried out in northern Burkina Faso. The technique used is based on linear correlation between millet yield and the time integral of the Normalized Difference Vegetation Index (iNDVI) derived from NOAA AVHRR data. In contrast to earlier related studies, the correlation has been established using satellite data extracted exclusively within the agricultural domain. The integration period for the iNDVI correponds to the reproductive phase only of the growing period of millet. Furthermore, iNDVI can also be used to estimate the acreage or the agricultural domain, by the application of a suitable threshold to classify areas into agricultural and non-agricultural domains.

It is therefore possible to assess the yield and the acreage of the agricultural domain and to derive an estimate of the millet production of the area by the end of the season, on the basis of NOAA AVHRR data alone.     

Assessment of ecological conditions associated with the 1980/81 desert locust plague upsurge in West Africa using environmental satellite data

National Oceanic and Atmospheric Administration (NOAA) satellite data from the Advanced Very High Resolution Radiometer (AVHRR) sensor were analysed to document the vegetation biomass dynamics associated with the regional desert-locust upsurge in West Africa during 1980/81, which affected an area of some 600 000 km² in Mali, Niger and Algeria. Comparisons were made among locust population survey reports, rainfall records from eighteen stations in the same area, and the satellite data in vegetation index format. The satellite-recorded temporal and spatial distributions of desert vegetation biomass were closely correlated with both the locust population surveys and the available rainfall data. An attempt was made to develop a quantitative relationship between a satellite-derived potential breeding activity factor (PBAF) and the observed desert locust populations. Analysis of the multitemporal satellite data set indicates that, had the NOAA/AVHRR vegetation index data been operationally available in June 1980, effective preventive control measures would have only been necessary for an area of 600 km².     

A Review of: “Environmental Information Management and Analysis: Ecosystem 10 global scales”. Edited by W. K. MICHENER,J. W. BRUNT and S. G. STAFFORD (London: Taylor &: Francis Ltd, 1994). [Pp. 535.] Price £55,00.

Abstract

Satellite indices of vegetation from the Australian continent were calculated from May 1986 to April 1987 from NOAA-9 AVHRR (Advanced Very High Resolution Radiometer) and Nimbus-7 SMMR (Scanning Multichannel Microwave Radiometer) satellite data. The visible (VIS) and near infrared (N1R) reflectances and their combination, the Normalized Difference (ND) Vegetation Index were calculated from the AVHRR sensor. From the SMMR, the microwave Polarization Difference (PD) was calculated as the difference between the vertically and horizontally polarized brightness temperatures at 37 GHz. The AVHRR data were gridded to match the 25 km spatial resolution of the SMMR 37 GHz data and both data sets were analysed to provide a temporal resolution of one month. Using a one month lag, the ND, PD, VIS and NIR, indices were plotted against rainfall and water balance estimates of evaporation, calculated using the monthly rainfall data and long term averages of pan evaporation from 74 locations covering a range of vegetation types. The monthly plots had wide scatter. This scatter was reduced markedly by aggregating the data over twelve months, leading to the conclusion that direct satellite monitoring of annual evaporation across the Australian continent using PD or VIS is feasible for areas with evaporation less than 600 mm y^?1. The ND relationship was limited by scatter and the PD and VIS relationships by their saturation above 600 mm y^?1, which spanned about two-thirds of the continental range studied. Scatter was reduced and ND had a predictive range above 600 mm y^?1 if evaporation was normalized by evaporative demand. But prior knowledge of potential evaporation is needed in this approach. The NIR reflectance of forests were consistently lower than neighbouring areas of agriculture, thus ND may underpredict the evaporation of forests relative to agriculture. Temporal resolution of the satellite indices over periods of one month could not be evaluated due to spatial and temporal variability of climatic and biological factors not accounted for in the water balance estimates of evaporation.     

Comparison of precipitable water over Hawaii using AVHRR-based split-window techniques,GPS and radiosondes

Precipitable water vapour (PWV) was estimated over Lihue, Kauai, from AVHRR data using split-window techniques. The predicted values using the satellite sensor data were compared to precipitable water vapour amounts obtained from radiosondes and corrected GPS measurements. Compared to the corrected GPS precipitable water, the Dalu and RV satellite methods had rms errors of 7.3 and 3.8 mm, respectively. Typical values of PWV over Hawaii are approximately 27.5 mm, suggesting errors of about 14% in values estimated using the satellite split window technique near Hawaii.     

基于AVHRR和TM数据的时间序列较高分辨率NDVI数据集重构方法

由于技术条件的限制,一个传感器很难同时具有高空间分辨率和高时间分辨率。然而,在高分辨率尺度上监测地表景观季节性变化的能力是全球的迫切需要,融合周期短、覆盖范围广与分辨率高、周期长的遥感数据是一种较好的方法。基于AVHRR时间分辨率高和TM空间分辨率高及其数据积累时间长的特点,选择若尔盖高原为研究区域,在改进ESTARFM方法的基础上,对TM NDVI和AVHRR NDVI进行融合,构建高时空分辨率的NDVI数据集。研究结果表明:该方法能有机结合AVHRR NDVI的时间变化信息与TM NDVI的空间差异信息,有效实现高时空分辨率NDVI数据集的重构,3景预测高分辨率NDVI与MODIS NDVI产品相关系数分别达到了0.89、0.91和0.85。该方法能够在时间上保留高时间分辨率数据的时间变化信息,同时在空间上反映高空间分辨率数据的空间差异信息,从而为有效构建相对高分辨率时间序列NDVI数据集提供了可能的方法。     

Satellite and ground-based pasture production assessment in Niger: 1986-1988

Abstract

The standing crop of herbaceous biomass produced during the 2-4?month summer rainy season by the annual grasses in the Sahel zone provides an indication of resource availability for livestock for the following 9-month dry season. Combined use of NOAA advanced very high resolution radiometer (AVHRR) local area coverage (LAC) satellite data and biomass data, obtained through vegetation sampling of 25-100 km₂ areas, allowed the development of a method for biomass assessment in Niger. Vegetation sampling involved both visual estimates and clipped plots (double sampling). The relationship between time-integrated normalized difference vegetation index (NDVI) statistics derived from NOAA AVHRR LAC data (dependent variable) and total herbaceous biomass (independent variable) was obtained through regression analysis. An inverse prediction was used to estimate biomass from the satellite data. Biomass maps and statistics of the grasslands were produced for the end of each rainy season: 1986, 1987 and 1988. This information is being used for planning purposes by the pastoral resource managers of the Government of Niger.     

Geostationary satellite detection of bio mass burning in South America

Abstract

This paper presents the results of using Geostationary Operational Environmental Satellite (GOES) Visible Infrared Spin Scan Radiometer Atmospheric Sounder (VAS) data to monitor biomass burning associated with deforestation and grassland management in South America. The technique of Matson and Dozier has been adapted to GOES VAS short-wave and long-wave infrared window data to determine ihe size and temperature of fires associated with these activities. Although VAS data do not offer the spatial resolution available with Advanced Very High Resolution Radiometer (AVHRR) data (7 km versus I km) this decreased resolution does not seem to hinder the ability of the VAS instrument to delect fires; in some cases it proves to be advantageous, in that saturation does not occur as often. Sequences of VAS visible data are helpful in verifying that the hot spots sensed in the infrared are actually related to fires. Furthermore, the smoke of the fires can be tracked in time to determine their motion and trajectory. In this way, the GOES satellite offers a unique ability to monitor diurnal variations in fire activity and transport of related aerosols.     

The GBA2000 initiative: Developing a global burnt area database from SPOT-VEGETATION imagery

The scientific community dealing with modelling of emissions of greenhouse gases and aerosols from anthropogenic sources demands reliable and quantitative information on the magnitude of biomass burning at a global scale. It is in this context that the Global Burnt Area -- 2000 (GBA2000) initiative has been launched. The specific objectives of this initiative are to produce a map of the areas burnt globally for the year 2000, using the medium resolution (1.1 km) Système Pour l'Observation de la Terre (SPOT) 4-VEGETATION (SPOT-VGT) satellite imagery and to derive statistics of area burnt per country, per month and per main type of vegetation cover. A series of regional algorithms has been developed and incorporated into a data processing system designed to yield monthly estimates of areas burnt at a global scale. The map data will then be transformed into quantitative information and made publicly available over the World Wide Web at a range of spatial and temporal resolutions to satisfy some of the requirements of the atmospheric and climate change modelling community.     

Spatial structure,sampling design and scale in remotely-sensed imagery of a California savanna woodland

Abstract

This article describes research related to sampling techniques for establishing linear relations between land surface parameters and remotely-sensed data. Predictive relations are estimated between percentage tree cover in a savanna environment and a normalized difference vegetation index (NDVI) derived from the Thematic Mapper sensor. Spatial autocorrelation in original measurements and regression residuals is examined using semi-variogram analysis at several spatial resolutions. Sampling schemes are then tested to examine the effects of autocorrelation on predictive linear models in cases of small sample sizes. Regression models between image and ground data are affected by the spatial resolution of analysis. Reducing the influence of spatial autocorrelation by enforcing minimum distances between samples may also improve empirical models which relate ground parameters to satellite data.     

Evaluation of MODIS and NOAA AVHRR vegetation indices with in situ measurements in a semi‐arid environment

Much effort has been made in recent years to improve the spectral and spatial resolution of satellite sensors to develop improved vegetation indices reflecting surface conditions. In this study satellite vegetation indices from the Moderate Resolution Imaging Spectroradiometer (MODIS) and the Advanced Very High Resolution Radiometer (AVHRR) are evaluated against two years of in situ measurements of vegetation indices in Senegal. The in situ measurements are obtained using four masts equipped with self‐registrating multispectral radiometers designed for the same wavelengths as the satellite sensor channels. In situ measurements of the MODIS Normalized Difference Vegetation Index (NDVI) and AVHRR NDVI are equally sensitive to vegetation; however, the MODIS NDVI is consistently higher than the AVHRR NDVI. The MODIS Enhanced Vegetation Index (EVI) proved more sensitive to dense vegetation than both AVHRR NDVI and MODIS NDVI. EVI and NDVI based on the MODIS 16‐day constrained view angle maximum value composite (CV‐MVC) product captured the seasonal dynamics of the field observations satisfactorily but a standard 16‐day MVC product estimated from the daily MODIS surface reflectance data without view angle constraints yielded higher correlations between the satellite indices and field measurements (R ² values ranging from 0.74 to 0.98). The standard MVC regressions furthermore approach a 1?:?1 line with in situ measured values compared to the CV‐MVC regressions. The 16‐day MVC AVHRR data did not satisfactorily reflect the variation in the in situ data. Seasonal variation in the in situ measurements is captured reasonably with R ² values of 0.75 in 2001 and 0.64 in 2002, but the dynamic range of the AVHRR satellite data is very low—about a third to a half of the values from in situ measurements. Consequently the in situ vegetation indices were emulated much better by the MODIS indices than by the AVHRR NDVI.     

Statistical models of fragmented land cover and the effect of coarse spatial resolution on the estimation of area with satellite sensor imagery

The feasibility of correcting for errors in apparent extent of land cover types on coarse spatial resolution satellite imagery was analysed using a modelling approach. The size distributions for small burn scars mapped with two Landsat Multi-spectral Scanner (MSS) images and ponds mapped with an ERS-1 synthetic aperture radar (SAR) image were measured using geographical information system (GIS) software. Regression analysis showed that these size distributions could be modelled with two types of statistical distributions a power distribution and an exponential distribution. A comparison of the size distributions of small burn scars as observed with the Landsat MSS imagery to the distribution observed with National Oceanic and Atmospheric Administration (NOAA) Advanced Very High Resolution Radiometer (AVHRR) imagery indicated that distortions due to the coarse spatial resolution of AVHRR caused overestimation of the burn area. This bias was primarily caused by detection in two or three AVHRR pixels of burns whose actual size was on the order of a single AVHRR pixel. Knowledge of the type of the actual size distribution of small fragments in a scene and the causes of distortion may lead to methods for correcting area estimates involving models of the size distribution observed with coarse imagery and requiring little or no recourse to fine scale data.     

Combining AVHRR and ATSR satellite sensor data for operational boreal forest fire detection

The potential to combine data from two different satellite systems was studied to increase fire detection sensitivity and image acquisition frequency in real-time fire detection and fire control. A fully automatic fire detection algorithm was applied to all scenes that were acquired using both satellite systems. Local fire authorities were notified about each detected fire in their territory using real-time fire reports that were sent by telefax. The average time from the start of National Oceanic and Atmospheric Administration (NOAA), Advanced Very High Resolution Radiometer (AVHRR) image acquisition until the sending of a telefax fire report was 25 min. During the straw-burning season in April 2000, the Along Track Scanning Radiometer (ATSR) instrument detected twice as many fires as the AVHRR per unit image area. The main reason for this may be the average resolution cell of the ATSR, which is half the size of that of the AVHRR in terms of area. The response from fire authorities was used to estimate the number of correct alerts and false alarms. A false alarm rate of 12% and 7% was obtained in the fire seasons of 1999 and 2000, respectively.