EUMETSAT has developed a network of Satellite Application Facilities (SAF) for the future Application Ground Segments for the new generation Meteosat Second Generation (MSG) and European Polar System (EPS) platforms. Our main concern in LSA SAF is to develop an operational algorithm for retrieving vegetation parameters. In particular, fractional vegetation cover (FVC) and leaf area index (LAI), which are key parameters in the description of both land-surface processes and land-atmosphere interactions. The LSA SAF vegetation products will be provided over the full MSG disk at 3-km spatial resolution with a temporal resolution of 10-days. The use of BRDF models assures that these products will be corrected of the surface anisotropy effects. The algorithm is based on the complementary use of variable and multiple endmember spectral mixture analysis (DISMA), according with the available directional sampling. Land cover map, soil type databases and the clumping index are auxiliary information in the prototype. The prototyping algorithm has been tested using both airborne POLDER data over croplands, and the POLDEr on ADEOS BRDF database. A first version of the prototype for the MSG developed on synthetic MSG data is already implemented in the LSA SAF system. In this paper, the prototyping algorithm designed to retrieve the LSA SAF vegetation products and its validation on the above mentioned data sets are presented.
The availability of quasi-simultaneous multi-directional measurements from space, as provided by the Multiangle Imaging SpectroRadiometer (MISR) on board the Terra platform, offers new and unique opportunities to document the anisotropy of terrestrial surfaces at key solar wavelengths. This contribution outlines the physical reasoning underpinning a new quantitative interpretation of multi-angular reflectance measurements over terrestrial surfaces. The most innovative aspect of this approach concerns the characterization of the heterogeneity of these surfaces. Indeed, when appropriately parameterized, the shape of the reflectance anisotropy at specific optical wavelengths can be related to the structural characteristics of the observed target. This allows the detection of geophysical conditions for which surface heterogeneity is an essential ingredient to describe the measured reflectance pattern. This finding paves the way for the quantitative characterization of plant canopy structure on the basis of multi-angular data.
The operational processing of NOAA-AVHRR data and the derivation of vegetation index (NDVI), leaf area index (LAI) and vegetation cover fraction for the European Alps is presented. The analysis was done for three elevation zones (<500m, 1000-1500m and >2500m) to show the dynamic characteristic of vegetation in the years 1995 to 1998. The vegetation cover fraction shows a high variability in lower elevations during winter caused by the not persistent snow cover. In elevations above 2500m the high variability could be detected during summer. The exponential approach to derive LAI using NDVI data is only valid for elevations above 2000m or for NDVI less than 0.5. Otherwise the LAI values are saturated because small changes in NDVI result in an increased range of LAI up to 1.5. This prevents an exact
derivation of leaf area index based on the normalized difference vegetation index.
In this work we present an innovative method for retrieving vegetation variables whilst at the same time making optimal use of the new generation satellite sensors. The approach is aimed to the generation of vegetation products exploding the angular capabilities provided by the MSG/SEVIRI and EPS/AVHRR within the LSA SAF Project. The products include leaf area index (LAI) and fractional vegetation cover (FVC). The algorithm is based on the complementary use of Variable Multiple Endmember Spectral Mixture Analysis (VMESMA) and the inversion of a light-canopy interaction model, namely DISMA (DIrectional Spectral Mixture Analysis), which combines the geometric optics of large scale canopy structure with principles of radiative transfer for volume scattering within individual crowns. Unlike VMESMA, DISMA fully accounts for additional information on directional anisotropy. The prototype has been implemented in the LSA SAF system and tested using SEVIRI synthetic data. The algorithm validation includes feasibility analyses, sensitivity assessments as well as evaluation of the prototype on SEVIRI synthetic data. The study contributes to assess the uncertainties with SEVIRI based vegetation products.
The use of remote sensing data in site specific crop management aims at the prediction of soil and crop factors that have an impact on yield formation processes in agriculture. Numerous methods demonstrate the potential of spectral reflectance data for the detection of qualitative and quantitative crop features but there is, however, no established methodology for the implementation of these data in operational crop production processes. The paper describes the main aspects of remote sensing based site characterization, considering major site variables (yield, soil) and plant parameters (nitrogen uptake) as key features for the description of the site specific variability in crops. Spectral reflectance data of the VIS/NIR region are transformed into different spectral indices for statistical analysis. Analyzing these indices it is found that the determination of a prediction model depends on the relevance of the suggested data fitting method (causality) as well as on the statistical significance of the interrelationship. Results point out that remote sensing data are suitable predictors for crop vitality and site characterization. Hence, the application of these data in agricultural work routines is limited by their quality and availability as well as by the influence of environmental factors on yield formation processes.
AVIRIS data from Blythe were acquired in June 1997 to study the agricultural spectra from different crops and for identification of crops in other areas with similar environmental factors and similar spectral properties. The main objectives of this study are: 1) to compare the spectral and radiometric characteristics of AVIRIS data from verities of cotton crop with the spectra measured by FieldSpecR ASD radiometer; 2) to explore the use of AVIRIS images in identifying agricultural crops; 3) to study the impact of environmental factors on selected crops and; 4) to build a spectral library for the cotton crop varieties that were studied. A long-term goal is to extend the spectral library for different vegetation or crops in different stages of growth or different varieties. In order to support our study, on June 26, 2001 we collected spectral data using the FieldSpec spectrometer from selected fields planted with different cotton varieties at Blythe area, California (at the Longitude 114° 41.88 W and Latitude 33° 24.27N to Longitude 114° 41.86 W and Latitude 33° 24.00N). The spectral data of cotton varieties were studied. Environmental parameters were studied such as the soil water content (WC), pH, organic matter (OM), C% and nitrogen (N%). The results of this study showed that there were differences in the signatures of different cotton varieties. Also, there was a significant correlation between the data that were collected by AVIRIS image scene in 1997 and spectral data collected by the FieldSpec spectrometer. This correlation allowed us to build a spectral library to be used in ENVI-IDL software. This leads to identification of different cotton varieties and in particular the visible part of the spectra. AVIRIS data are in agreement with FieldSpec data. Using IDL algorithms of Spectral Angle Mapper classification (SAM), Spectral Feature Fitting (SFF) and Spectral Binary Encoding (SPE) showed that there is an excellent agreement between the predicted and the actual crop type (i.e. the correlation is between 85 - 90% match). Further use of the AVIRIS images can be of a value to crops identification or crops yield for commercial use. The yield data of Cotton varieties were correlated significantly with the spectral data from the AVIRIS and from the hand-held radiometer and it showed the impact of different environmental parameters on the yield of the crop.
Author(s): Mieke Reyniers; Els Vrindts; Josse de Baerdemaeker; Pol Darius
What is lacking in precision farming at present are more comprehensive and fast non-destructive methods for obtaining the data needed to prescribe varia*ble treatments. In precision farming there is a demand for sensors that can easily monitor crop nitrogen requirements throughout the growing season on a high resolution. Currently used optical measurement platforms like satellites, airplanes and hand-held sensors, do not meet the needs of precision agriculture for good nitrogen management possibilities. An automated sensor system mounted on a tractor was developed and used to detect crop nitrogen status optically. A line spectrograph was used to detect amount of nitrogen (kgN/ha) and chlorophyll (kg/ha) in a wheat crop (Triticum aestivum L.). By calculating the red edge inflection point of the plant spectra in the images, wheat crop nitrogen stress within small areas in the field could be detected. Spectrograph red edge was highly correlated with applied nitrogen to the wheat crop (0.90), with crop nitrogen uptake (0.89) and with chlorophyll amount in the crop (0.86). The average errors when estimating those variables with the red edge inflection point were -0.4% (24.15kgN/ha), -1% (17.25kgN/ha) and -10% (14.74kg/ha) respectively. This means that spectrograph red edge measurements of the wheat crop during the growing season can be a predictor of topdress nitrogen needs.
It is shown that the spectral curve of reflectance of vegetation
contains the sufficient information to create a set of parameters
for effective monitoring of agricultural crops. Most of them are
based on the chlorophyll estimation or characteristics, which are
dependent on specific influence of inner structure of plant
tissues on leaf reflection in the region of chlorophyll
absorption. New chlorophyll indices are proposed for estimation of
chlorophyll content in leaves using the shape of leaf reflectance
curves. The ratio of two maxima in the 1-st derivative plot from
reflectance spectral curve in 680-750 nm region has been shown
to correlate with chlorophyll content in winter wheat leaves.
Independent component analysis of reflectance spectral curves has
been applied as well. An interrelation between the chlorophyll
concentration and vectors of principal components has been found.
The estimates of the chlorophyll content by using of these
parameters and regression equations gave suitable results.
Comparison of two approaches has been performed. Stability of both
approaches with regard to incomplete project covering have been
tested. Usage of physical and graphical models permits to estimate
stability in calculation results of chlorophyll concentration
influence of soil reflection. It has been shown that the ratio of
two maxima in the 1-st derivative plot was changed now more than
5% and 11% under 50 % and 25 % projective cover,
respectively, on a background of dark soil or sand. The
reflectance coefficient at 550 nm correlates with chlorophyll
content but it is highly sensitive to contribution of soil
reflectance. Therefore combination of chlorophyll estimates
obtained by red edge parameters and the reflectance coefficient at
550 nm gives possibility to estimate a projective covering. We
shown that principal components approach is resistant to
influence of project covering.
Crop coefficients (Kc) are defined as the ratio of actual crop evapotranspiration to the evapotranspiration (Eto) of a grass reference crop, often taken from Penman/Monteith’s methodology. They are used to estimate theoretical crop evapotranspiration (Etc). Actual evapotranspiration is measured in field lysimeters but this lacks acceptability when applied to the field or large irrigation schemes where conditions are very variable. This paper uses the SEBAL method (Bastiaanssen et al., 1998) to determine actual evapotranspiration from satellite images. This is compared with estimated Eto from both Penman/Monteiths methodology and Eto estimated from a SEBAL estimation of reference crop evaporation within a project. The range of Kc values for the cotton crop were then calculated. Crop coefficient Kc maps were made for two irrigation projects. The methodology was applied on a large cotton irrigation scheme (Starikan) southern Kazakhstan. The estimated mean real time Kc value of 1.16 (±1-3% error) was higher than the standard Kc value of 1.1 identified by FAO-56. The methodology was validated on a different date and different irrigation schemes (Chardara) 200 km south Starikan. The methodology is discussed.
A linear relationship between NDVI and basal crop coefficient (Kcb) allows to compute the spectral crop coefficient (Krcb). Due to the influence of soil variations varying surface humidity on NDVI, five soil optimized indices have been used to obtain a linear relationship normalized for soil background effect (SAVI, OSAVI, TSAVI, MSAVI and
GESAVI). Data used on this work have been obtained from a field campaign for corn in the area of Barrax (Spain), describing crop growth stages with green fraction cover (GFC), and leaf area index (LAI). SAVI with optimized factor L set to 0.5 is a good estimator of Krcb from sparse to dense vegetation, nevertheless the soil line based index ( GESAVI) due to a wider range of variation are more sensitive to leaf variations at high levels of vegetation amount. Spectral crop coefficients obtained from SAVI and soil line based GESAVI are sensitive to crop hazards by weather anomalies and
estimates in real time the basal crop coefficients to estimate the amount of water removed by the crop from the active root zone.
With the aim to derive crop water requirements (ETp) for an irrigated area covered by orange orchard in Sicily, Quick Bird and ASTER TERRA high resolution satellites data were used and compared with reference to their different spatial and spectral resolution. Satellites data allowed to improve the monitoring of canopy development in the irrigated area by identifying biophysical vegetation variable (LAI, albedo, vegetation indicators, etc); this information was successively used for the evaluation of maximum crop water needs by means of the well known Penman-Monteith equation. The paper results evidence the importance of very-high resolution sensors such as QuickBird in areas characterised by strong spatial heterogeneity. The algorithms applied to estimate the canopy parameters and the crop water requirements were applied by considering different levels of radiometric calibration of the satellite data, which produced marked
differences in the final results.
Several satellite sensor systems useful in Earth observation and monitoring have recently been launched and their derived products are being used in regional and global vegetation studies. The joint use of these multi-resolution sensors offers many opportunities for vegetation studies. Spectral vegetation indices obtained from Landsat, Spot, IRS and other sensors are now widely available for monitoring ecosystem dynamics. However, the joint use of data from different satellites requires inter-satellite cross-calibration. We will use a multi-temporal data synthesising procedure for this purpose.
In this paper we analyze the broadband reflectance and NDVI relationships among the various relevant sensors. The key to the method is in using synchronous or nearsynchronous imagery from different sensors.
Comparison between reflectances for different bands shows that a linear function fits well to describe the relation between different sensors. Observations made from different sensors at different spatial scales can be reliably compared only if they are spatially aggregated to an adequate grid size. This minimum spatial aggregation size depends on the spatial resolution of the sensors involved in the comparison. In any case, it must be at least 3 x 3 pixels of the coarser resolution sensor.
Author(s): Junbang Wang; Zheng Niu; Binmin Hu; Changyao Wang
The determination of terrestrial ecosystem carbon source/sink spatial pattern is becoming one of the hottest problems and many environment politics focus on it. As a new tool for terrestrial ecosystem carbon modelling at large scale from field plot, to region, to global, remote sensing is applied to initialize, drive, and validate the model, combined with geophysics information system (GIS) and computer modelling. Carbon flux models with remote sensing data as input may be classified as light use efficiency model, process model, and eco-physiological model based on “big leaf” hypothesis. The model generally includes two parts: NPP and soil respiration model to estimate carbon flux based on the principle that the carbon flux of ecosystem equal NPP minus heterogeneity respiration (soil respiration).
Remote sensing, however, is more applied in NPP modeling but little in soil respiration estimation. The latter mostly based on relationship between soil respiration and soil temperature and is highly developed. Since remote sensing is applied to retrieve land surface temperature (LST) with infrared waveband, a hypothesis was put forward, that is, land surface temperature retrieved from infrared waveband can substitute soil temperature to estimate soil respiration. The hypothesis was validated with a field experiment and result was given in this article.
The experiment located in a winter wheat field at Quzhou experiment station, Hebei province, China, from Apr 19 to May 20, in 2002. The soil respiration rate was measured with CID photosynthesis system, and canopy infrared temperature, soil surface temperature were measured respectively at same time. The station provided us soil moisture content data of whole growth of winter wheat. The result shows that the soil CO2 efflux from winter wheat field is -0.03~1.38μmolm-2s-1. Its diurnal variation is well fitted with univariate quartic curve. Its variation in winter wheat heading growth period well coincide with temperature and soil moisture content. The Pearson correlation analysis shows that, on the averaged sense, for a day, soil CO2 efflux significantly correlated with the temperature of the air (Tair), the soil surface (Tsur), the averaged thermal (Tinf) temperature respectively at the p-level<0.001.
The relation between soil respiration and canopy thermal temperature (Tinf) and soil surface temperature (Tsur) was modeled with equations from Fang and Moncrieff (2001) respectively. On the whole, the performance of models with Tinf as independent is better than one with Tsur as independent for the data on May 8. The max multiple correlation coefficient (MCC) of the former is 0.95118 large than the MCC 0.92338 of the later, which provide a better fundament for the hypothesis above. The result of model analysis shows that the one of Schlentner and Van Cleve (1985) is best candidate in this study because of its high coefficient of determination and its principle.
However some problem should be improved in the future. Firstly, soil respiration was measured with CID photosynthesis system and chamber which demand to consider the disturbance of chamber and the precision of the instrument. Secondly, the research focus on a point not on whole area comparing with the resolution remote sensing image, such as NOAA/AVHRR, TM, MODIS, since the result can not be directly applied to satellite image, that is, the experiment on a large spatial scale should be done for satellite image application.
In this study an innovative approach for investigating the accumulated meteorological effects on cotton production during the growing season is presented. The quantification of the meteorological effects is based on the incorporation of the Bhalme and Mooley Drought Index (BMDI) methodology into the Vegetation Condition Index (VCI) extracted by NOAA/AVHRR data. The resulted Bhalme and Mooley Vegetation Condition Index (BMVCI) uses the same scale as the Z-Index of the Palmer Drought Severity Index (PDSI) for drought monitoring. The study area consists of the country of Greece. Eighteen years of NOAA/AVHRR data are examined and processed with the BMVCI to examine the unfavourable conditions for cotton production. For the validation of BMVCI an empirical relationship between the cotton production and the BMVCI values is derived. The method is developed based on the first sixteen years time series data and validated utilizing the following two years. The resultant high correlation coefficient and the approximation of the production for the validated years refer to very favourable results and confirms the usefulness of this integrated methodological approach as an effective tool to assess cotton production in Greece.
The use of remote sensing to control area based rural development measures is investigated in selected test sites in the three Member States France, Germany and Spain. Measure definition and implementation are checked if a remote sensing control is feasible. For selected measures very high resolution imagery from Quickbird, Ikonos and airborne photos are used to define image interpretation rules for an effective control process and a control design will be
According to the specific situations in China, this paper discusses the application of RS data in analysing the dynamic balance between cultivated land supply and demand which is one of main tasks of the Ministry of Land and Resources of China. It points out that based on applying RS data to monitoring land use changes, we can make full use of RS data to extract the information required in the analysis on the balance, which is an important approach for dynamically mastering and regulating the balance. It presents the framework and main aspects for analysing the balance, including the environment of the balance, the elements of the balance, the state of the balance and the process of the balance, as well as analysis on the balance at multimeasures, such as the balance in quality, in Gross Amount, in Per capita Amount, in Region and in Time.
This paper presents the state of the art of the general principle of liquid flow measurements by ultrasonic method, and problems of
flow measurements. We present an ultrasonic flowmeter designed according to smart sensors concept, for the measurement of irrigation
water flowing through pipelines or open channels, using the ultrasonic transit time approach. The new flowmeter works on the principle of measuring time delay differences between sound pulses transmitted upstream and downstream in the flowing liquid. The speed of sound in the flowing medium is eliminated as a variable because the flowrate calculations are based on the reciprocals of the transmission times. The transit time difference is digitally measured by means of a suitable, microprocessor controlled logic. This type of ultrasonic flowmeter will be widely used in industry and water management, it is well studied in this work, followed by some experimental results. For pressurized channels, we use one pair of ultrasonic transducer arranged in proper positions and directions of the pipe, in this case, to determine the liquid velocity, a real time on-line analysis taking account the geometries of the hydraulic system, is applied to the obtained ultrasonic data. In the open channels, we use a single or two pairs of ultrasonic emitter-receiver according to the desired performances.
Finally, the goals of this work consist in integrating the smart sensor into irrigation systems monitoring in order to evaluate potential advantages and demonstrate their performance, on the other hand, to understand and use ultrasonic approach for determining flow
characteristics and improving flow measurements by reducing errors caused by disturbances of the flow profiles.
The objective of this study was to process multitemporal satellite data in order to detect burnt areas and classify these areas according to how many times they have been burnt. The area of study is situated in Western Peloponnese near the site of Ancient Olympia. In 1986, 1998 and 2000 three big fires have burnt more than 500.000.000 m2 of forest and rural land in the broader area.
In order to detect the vegetation changes and classify the burnt areas for the period 1984-1999 we used the following multitemporal satellite images:
A Landsat 5 TM cloud free subscene, acquired on July 27 1984,
A Landsat 5 TM cloud free subscene, acquired on September 18 1986,
A Landsat 7 ETM cloud free subscene, acquired on July 28 1999,
We applied the NDVI (Normalized Difference Vegetation Index) to all the satellite images. Then, we created two new images with two bands each. one using the vegetation indexes images of 1986 and 1984 and a second one using the vegetation indexes images of 1999 and 1986. Then, we applied the PCA method to the new images.
After the fires of 1986, 1998 and 2000 local authorities have mapped the burnt areas using traditional methods. With joint use of the thematic maps and the above produced images of Principal Components we managed to classify the burnt areas according to how many times the have been burnt.
The general conclusion is that we can use satellite data with the vegetation indexes PCA method for the accurate mapping of burnt areas and the vegetation monitoring. Burnt areas for more than twice cannot be regenerated on its own so the classification of the burnt areas according to how many times they have been burnt is very important in order to locate the areas that needs reforestation.
Increasing concern about environment and interest to avoid losses led to growing demands on space borne fire detection, monitoring and quantitative parameter estimation of wildfires. The global change research community intends to quantify the amount of gaseous and particulate matter emitted from vegetation fires, peat fires and coal seam fires. The DLR Institute of Space Sensor Technology and Planetary Exploration (Berlin-Adlershof) developed a small satellite called BIRD (Bi-spectral Infrared Detection) which carries a sensor package specially designed for fire detection. BIRD was launched as a piggy-back satellite on October 22, 2001 with ISRO’s Polar Satellite Launch Vehicle (PSLV). It is circling the Earth on a polar and sun-synchronous orbit at an altitude of 572 km and it is providing unique data for detailed analysis of high temperature events on Earth surface. The BIRD sensor package is dedicated for high resolution and reliable fire recognition. Active fire analysis is possible in the sub-pixel domain. The leading channel for fire detection and monitoring is the MIR channel at 3.8 μm. The rejection of false alarms is based on procedures using MIR/NIR (Middle Infra Red/Near Infra Red) and MIR/TIR (Middle Infra Red/Thermal Infra Red) radiance ratio thresholds. Unique results of BIRD wildfire detection and analysis over fire prone regions in Australia and Asia will be presented. BIRD successfully demonstrates innovative fire recognition technology for small satellites which permit to retrieve quantitative characteristics of active burning wildfires, such as the equivalent fire temperature, fire area, radiative energy release, fire front length and fire front strength.
Over the years, several satellite-based indicators have been developed for Fire Susceptibility Estimation (FSE) most of which are based on AVHRR data. Unfortunately, these indicators yield different results when applied to different ecosystems or geographic regions and this creates confusion concerning their effectiveness. This work aims at evaluating the performance of the existing AVHRR-based FSE methods. Such evaluation was performed in the Basilicata and Calabria Region by using NOAA-12,-14 summery imagery selected from a long time series acquired from 1995 to 1999. Fire susceptibility maps
obtained from the considered methods were compared to fire archives provided by the Italian National Forestry Service. The most satisfactory results were obtained from methods based on the cross analyses of NDVI (Normalized Difference Vegetation Index) with thermal channels.
This study investigates the potential of classifying complex ecosystems by applying the radial basis function (RBF) neural network architecture, with an innovative training method, on multispectral very high spatial resolution satellite images. The performance of the classifier has been tested with different input parameters, window sizes and neural network complexities. The maximum accuracy achieved by the proposed classifier was 78%, outperforming maximum likelihood classification by 17%. Analysis showed that the selection of input parameters is vital for the success of the classifiers. On the other hand, the incorporation of textural analysis and/or modification of the window size do not affect the performance substantially. The new technique was applied to the area of Lake Kerkini (Greece), a wetland of great ecological value, included in the NATURA 2000 list of ecosystems.
Temporal changes of terrestrial vegetation have traditionally been monitored using empirical remote sensing tools, which are sensitive to perturbations as well as to the spectral properties of the sensor. Advances in the understanding of radiation transfer theory, and the availability of higher performance modern instruments, have led to the development of physically-based inverse methods to derive biogeophysical products. Jointly, these developments allow the retrieval of reliable, accurate information on the state and evolution of terrestrial environments. A series of optimized algorithms has been developed to document biogeophysical variables, and in particular to estimate the Fraction of Absorbed Photosynthetically Active Radiation (FAPAR) from a variety of optical instruments. As a result, monitoring managed (e.g., agriculture) or natural ecosystems will benefit from the availability of local, regional and global time series of remote sensing products such as FAPAR. This paper outlines the methodology and exhibits selected results in the form of temporal composites derived from the SeaWiFS sensor.
This paper summarizes our research on land use and cover changes in the critical areas named North Ningxia (Yinchuan Region), North Shaanxi (Yulin Prefecture/Mu Us) and Middle Tarim River in northwest China. The objectives of the study were to investigate the present land use situation and changes in the past decades and to understand causes of landuse changes. Multi-temporal Landsat TM and ETM+ images (plus an old Corona image for the Middle Tarim), countylevel
socioeconomic data and meteorological data were used for this task. The methods and procedures adopted in this work were image registration, atmospheric correction, tasseled cap transformation, indicator differencing, county-level change mapping, and multivariate regression modeling. The principal conclusions are as follows: (1) Not “advancing desert” was observed; however, signs of serious land degradation, e.g., vegetation degradation and soil salinization, have
apparently taken place in the past decades due to cultivation practices, land reclamation and grazing. Some of these changes can be traced back to land use policies. (2) Farmland extension is a remarkable rural environmental change in these sites and is associated with the increase in agricultural output. Taking up a small percentage of the total change, the urban extension is related to about 90% of the GDP growth and driven directly by the urban population and their socioeconomic activities. Some river courses have been narrowing, owing partly to climate variability but mainly to the overuse of water in agriculture.
This work approaches the study of Cellular Automata to the simulation of Satellite Remote Sensing images applied to modeling environment landscape dynamics. The images were collected by SPOT and Landsat-MSS from one forest in different times. After the geometric correction and images treatment a binary map will be formed by pixels that contain information about the forest existence. The main purpose is to predict in a geographic map what will happen with the landscape forest in the future. The simulation is done through the analysis of the temporal maps in accordance with their progression, regression or stability in time and with rules that describes how CA do the
simulation. The results achieved are predict maps very useful for a environmental analysis. The experimental tests have showed promising results for studies related to forestry modeling.
Application of satellite remote sensing imageries in studying desert ecosystems is crucial normally due to desert extend, its harsh environment and difficult access which make studying and monitoring of desert ecosystems a cumbersome task.
Black Saxaul (Haloxylon Aphyllum), as a resistant plant, is widely planted in Iran’s central deserts to prevent sand dune movement and protect arable lands, roads and buildings from sand debris.
An attempt was made to study the effect of Saxaul plantation in Kavir-e-Omrani, in reducing wind erosion and stabilizing sand dunes after 30 years of its plantation in an area of 31627 hectares.
Landsat 7 ETM+ imagery acquired on March, 2002 was used to study Saxaul community extent and its canopy cover percentage classes to be related to the field measurements of soil sedimentation depth along prevailing wind direction and through canopy cover percentage gradient of Saxaul community. It was therefore necessary to have canopy cover percentage classes of the community obtained via classification of the ETM images, and their derivative bands.
All ETM bands were registered to 1:50000 topographic maps of the area and 10 GCPs obtained by filed measurements using GPS. Correction was made on digitized maps using linear transformation and nearest neighbor method for resampling with RMS error of less than 8 meters. 53 sampling units of 90m by 90m were field checked and canopy cover percentage, density (number of Saxaul per unit area) and ground canopy cover of accompanying plants were measured. Soil samples of ground surface were obtained within each sampling units for lab analysis of soil texture, EC (electric conductivity) and ESP (exchangeable sodium percentage). Coordinates of the corners of sample units were recorded using GPS so that positional discrepancies of sample units were minimized. 16 different vegetation indices, including RVI, NDVI, SAVI, MSAVI and other indices were created. Mean DN values of all ETM bands (except panchromatic band) and 16 derivative images related to 53 sampling units were extracted for statistical analysis. Principal Component Analysis and Correlation Analysis showed no meaningful correlation between canopy cover percentage classes of Saxaul and DN values of all ETM bands as well as 16 derivative images. To examine capability of ETM bands and their 16 derivative images to differentiate between Saxaul sites having more than %75 of canopy cover and sites with bare soil, discrimination of the two sites were tested via Student T test. A series of unsupervised classification was also performed on FCC image with 2, 3, 16 and unlimited number of classes using ISODATA clustering method to find if Saxaul plantation could be classified as a distinct class.
Results showed no distinction between the two sites. Visual investigation of all images proved the statistical results.
Despite the fact that the image was acquired in a season with highest Saxaul greenness and LAI, It was found that ETM images are unable to detect Saxaul plant community. It seems, inability of ETM to discriminate Saxaul plant community from surrounding bare soil is due to Saxaul prevailing bark percentage comparing to its LAI, its reduced leaf surface area as well as its pubescent leaf structure which seems to let beneath soil reflectance prevail upper plant cover reflectance.
Kernel-based reclassification algorithm derives information on specific thematic classes on the basis of the frequency and spatial arrangement of land cover classes within a square kernel. This algorithm has been originally developed and validated for the urban environment. The present work investigates the potential of projecting this technique to the classification of very high spatial resolution satellite imagery of natural ecosystems. For that purpose a software tool has been developed. The output, apart from the reclassified image, includes a post-classification probability map which shows the areas where the kernel reclassification algorithm has given valid results. The software was tested on an IKONOS image of Lake Kerkini (Greece), a wetland of great ecological value, included in the NATURA 2000 list of ecosystems. The results show that the algorithm has responded successfully in most cases overcoming problems previously encountered by pixel-based classifiers, such as pixel noise.
Lake surface water temperature (LSWT) are operationally derived from the National Oceanic and Atmospheric Administration operated Advanced Very High Resolution Radiometer (NOAA - AVHRR) data using a nonlinear sea surface temperature (NLSST) algorithm. The adapted method has been widely examined with the bias of the algorithm around 0.5°C or better. Preliminary analysis shows good agreement between satellite derived LSWT and in - situ measurements at two different lakes. A comparison of LSWT at noon (mean local time) for three lakes is presented. Surface water temperature variations are dominating the annual cycle, however, the varying geospatial attributes of each lake result in specific surface temperature characteristics. Lakes located close to each other can display considerable differences in average surface temperatures by as much as 3°C. Knowledge of this fact gives new insights and possibilities for modeling local scale meteorological phenomena like heat flux, energy budget and evapotranspiration. Using operational satellite-derived lake surface temperature can also improve numerical weather prediction models on local scales.
Water hyacinth (Eicchornia crassipes (Mart.) Solms) is an invasive aquatic macrophyte that has infested the lake Victoria, East Africa, since the late 1980s. It has been associated with major negative economic and ecological impact of this important water resource in East Africa. Remote sensing technology has significant potential in mapping this fast growing floating weed, in a mostly inaccessible area for field measurements.
Our study site is the Winam Gulf, on the Kenyan part of the Lake, which has had the highest reported infestation in recent years. The paper describes a study to evaluate the ability of ETM+ multispectral imagery in mapping water hyacinth and associated macrophytes in the hyacinth infested Winam Gulf. By applying hyperspectral techniques on multispectral data, a spectral mixture analysis was undertaken using image-derived endmembers. The study was also an evaluation of an alternative way of acquiring emergent macrophytic endmembers in cases where limitations like lack of hyperspectral data, spectrometric measurements and spectral libraries exist.
The results demonstrate that whereas it is possible to discriminate and map the different spectral constituents, a spectral library of the endmembers under investigation would be required for positive identification, especially for macrophytes that are closely related spectrally, fast growing, have varying concentrations (density) spatially, and are non-static in nature.
Remote sensing methods make it possible to analyze and describe landscape changes. However, one can hardly acquire sufficient data for direct long-term analysis. Multiple sensors, geometric distortions, phenological phase differences, atmospheric conditions, different solar angles and many other effects cause inter-scene variability. Furthermore, the temporal distribution of available data sets is often inhomogeneous, which tends to amplify the above-mentioned problems. In our work, we propose a methodology to cope with these difficulties for long-term environmental monitoring and quantitative change detection. A complex approach was chosen with the objective of integrating different methods and disciplines (radiometric and geometric correction, classification, image segmentation and GIS analysis, among others) to extract the maximum of information from the available data. This methodology is presented
and tested on an interesting case study that deals the environmental effects of a barrage system in the northwestern part of Hungary.
The monitoring of short-term changes in structural characteristics of forests is important to understand mechanisms of vegetation loss that can be associated with deforestation, and illegal logging. These changes, however, should be differentiated from variations in vegetation activity due to interannual variability. Change detection based on thematic information is limited for this purpose because it depends highly on classification accuracies, and it does not allow a quantitative evaluation of biomass loss. The definition of bio-indicators associated with structural characteristics (such as, leaf area index, vegetation fraction) is at present, the only way to monitor such changes. We developed an evaluation system consisting of 4 bio-physical variables, estimated from visible red and near-infrared observations of the Enhanced Thematic Mapper, to monitor changes in forest biomass. The system is based in the application of algorithms to estimate leaf area index, the fractional vegetation cover, leaf vegetation index, and a sparse vegetation cover index, from radiometrically and atmospherically calibrated data. The algorithms were applied to individual scene images acquired during the dry season (April-May) to maximize the forest vegetation signal, and in order to identify areas of change due to changes in forest biomass rather than changes in understory vegetation conditions. The change detection analysis consisted in comparing pixel-by-pixel scenes of such variables, and the results indicated that changes in structural characteristics of forest can be monitored with Landsat-7, being leaf area index, and fractional vegetation cover the most significant in identifying changes along roadsides and population centers that indicate biomass extraction.
In Algeria, arid and semi-arid regions occupy over than 95% of whole territory. Forests in the semi arid zone constitutes a front face to the advance of the desert towards northern sides. Like in other regions of the world, deforestation phenomenon have a serious consequences on the fragile ecosystem. Severe continuous drought, fires, pasture, insects as well as the absence of a clear forest politics are so many factors that reduced forest areas in this country. However, the conservation of this patrimony must be a priority of any regional development project.
This paper describes an evaluating study of the deforestation impact on forests in the region of Djelfa situated in the Saharian Atlas using multitemporal satellite remote sensing data. In order to establish a forest change map, a methodology based on the comparison between normalized difference vegetation indexes (NDVI) generated from satellite images was adopted. For this purpose, a pair of Landsat and (ETM+) images acquired over the region on April 11th, 1987 and march 24th, 2001 have been used.
Until being processed, data used have been geometrically and atmospherically corrected. Then, an (NDVI) have been produced for each date. Resulting from compared (NDVI) image presents the forest change map in the study area. Radiometric values of resulting image have been regrouped into three classes according to change types as follow :
Increased radiometry = more active vegetation
Decreased radiometry = deterioration in vegetation activity
Non changed areas = Non changed
Investigations made on the terrain permitted to interpret many causes of detected evolutions. Regressive changes were considerable and demonstrates however, the degradation effect on the vegetation state. Some of regressed radiometry are related to forest fires that affected the region in 1994. Almost of regressive changes are due to a deterioration of vegetation caused by multiple factors. Drought, deceases, pasture and infection are considered among the factors permitting the reduction of vegetative cover in these regions. Increased (NDVI) tones concerned generally reforested areas and are due to the fact that young plantations still more dense during the period 1987-2001. Other progressed changes are associated to the presence of some clouds in 1987.
Despite the limits of adopted methodology, we can say that resulting change map reflects the impact of deforestation on the region. It remains a helpful document for decision makers to take the preventive measures at least to decrease forest degradation.
In a study conducted for the U.S. Army Corps of Engineers, Mississippi Valley Division (MVD), we used ArcGIS software to interpolate, analyze, and display spatially explicit data describing fish and physical habitat factors (bathymetry, current velocity, and substratum) associated with a dike notching project in Bondurant towhead secondary channel in the lower Mississippi River between River Miles 390 - 394. Data were collected throughout project areas using hydroacoustic equipment. We used ArcGIS to interpolate coverages of each physical habitat variable, which were then compared with fish distribution data to determine patterns of habitat association. After analyzing data from several locations, we concluded that bathymetry, water velocity, and substrate composition
were most variable in areas immediately behind dike notches. However, the habitat diversity associated with notches was limited throughout the remaining portion of each project location. Data collected from throughout the side channel were analyzed. Habitat diversity (i.e., bathymetry, current velocity, and substratum) was greatest in
areas of immediate proximity with the notched dike. However, the lack of pre-notching data precluded a direct quantification of how dike-notching activities changed habitat quality.
This paper presents initial results of a hydrological study and flood assessment of the River Fani catchment in the Mirdita region in the north of Albania. Special attention is paid to Rubik town and its immediate surroundings. In the study multi-source and multi-scale data (published topographic and geological maps and various remote sensing sensors such as the Landsat series, MODIS and ASTER) are used to develop a GIS decision support system that uses hydrological inputs from various hydrological modelling packages - particularly WMS. The GIS is not only used as a link and management system between various data layers and hydrologic model outputs but also as modelling tool itself in an attempt to provide solutions to various catchment conservation issues, and flood prevention and control measure for Rubik town.
Preliminary results of change detection analysis using satellite images show high levels of reduced vegetation and increased urbanisation from 1984 to 2000. Changes in the River Fani channel morphology are mapped and analysed to areas of severe erosion, landslides and high flood risks.
It is concluded that remote sensing and GIS, when used creatively, provide powerful tools for overcoming the hydrological data challenge faced by hydrologists working on remote catchments with little or no structured conventional data acquisition system.
This work describes the validation of a distributed model for estimating direct recharge and evapotranspiration over arid and semiarid regions. This validation was performed for a lysimeter-site planted to festuca (grown under controlled irrigated treatment) and for two months, June and July 2003. The model, which can be classified as a distributed water balance model, puts its emphasis on two devising aspects. First, a detailed description of the effect of the land use on the water balance through processes of evaporation/transpiration and the evolution in time of the vegetated surfaces on the area. Second, the operational character of the model. The model was conceived to run integrated into a Geographical Information System and incorporates the pre-processing of the needed input parameters. This pre-processing comprises the use of remote sensing observations to monitor the plants status and their dynamics. In this study, agrometeorogical station records and information on irrigation scheduling, soil hydraulic properties and the festuca culture were used to run the model, whereas lysimeter measurements were used as validation data. Moreover, the performance of the model was checked for contrasting water conditions of the soil: completely wet and dried out.
Net radiation received at ground, is an important component of the energy-budget of ground surface. Indirect derivations of this parameter are common on the absence of direct measurements with reliable instruments. Most frequently, they are based partly on satellites observations and partly on available climatic observations. This paper presents development work on a tool for modelling the net radiation received at ground. This work presents a method to estimate both the solar (shortwave) and terrestrial (longwave) radiation at ground surface. At first, this work is devoted to estimate the shortwave irradiance under all sky conditions from the geo-stationary weather satellite Meteosat-7. The model is based on NREL’s SPECTRAL2 model for clear skies which takes no account of cloud attenuation of solar radiation. It can be shown to work well for its design purpose, but in maritime climates clear weather days are few and far between. Development work on the original model includes modification of the aerosol model and extended work on the Dedieu (1987) for developing a physically model to derive downward solar irradiance at the surface of the earth and surface albedo from Meteosat satellite measurements in the wavelength between 0.40 and 1.10 μm. The model takes into account Rayleigh and Mie scattering, water vapor and Ozone absorption. No threshold setting is necessary to distinguish between clear and cloudy conditions, there by avoiding the problem of its arbitrary nature and to some extent allowing quicker and easier data processing. Secondly, we propose an estimating net long wave radiation by available ground measurements from air and soil temperature combining Meteosat-7 and NOAA-14 AVHRR images.
Flood, is a kind of regional easily-occur natural calamity. It is important that people utilize non-structural measures for disaster-prevention and disaster-salvation. The pilot plant lies in Binjiang River Catchments, which is a branch of Zhujiang River (Pearl River) Guangdong Province China. The information will be extracted from remotely sensed TM images to plan flood risk area, and integrate meteorological data to divide the pilot plant into sections, quantify the flood risk degree to lives, belongings and environment in every sub-area. At last we will make a flood risk-planning map.
AVHRR (Advanced Very High Resolution Radiometer on board NOAA satellites) data are considered here to evaluate the possibility of using the surface temperature as an indicator of the soil/canopy water content at the short time-scale. This is obtained by means of an indirect approach based on a simplified soil-atmosphere energy balance. The techniques provide sufficiently detailed coverage of the processes in terms of the time and spatial scale with respect to hydrological applications. Two different approaches have been tried: the first based on thermal inertia measurements (Xue & Cracknell, 1995)1 through ATI (Apparent Thermal Inertia), the second based on surface temperature (LST) and vegetation indices (NDVI), with the TVDI (Temperature Vegetation Dryness Index) suggested by Sandholt et al. (2002)2. Both techniques were used in detecting moist areas in a single image (or day/night images), and in multitemporal multitemporal applications. In particular, a new cloud detection algorithm, based on the bimodal frequency distribution of the infrared brightness temperatures (Ch 5 AVHRR) when clouds affect the image, has been proposed for ATI. As regards TVDI, a modified technique has been proposed for fixing the warm edge of the triangle based on the detection of the extreme dryness conditions on a monthly basis. The modified TVDI has been tested in comparison with an antecedent precipitation index (API) for moisture detection in single images. The substitution of day/night land surface temperature differences instead of noon temperatures in the "triangle method" has been also tested with good results in the multitemporal approach. Application of the proposed techniques can allow one to track the evolution of soil moisture in space and time and to improve the knowledge on the relationship between vegetation (NDVI) and soil moisture dynamics.
The Soil Moisture and Ocean Salinity Mission is currently in the design and development phase (Phase-B) and scheduled for launch in early 2007. SMOS will exploit an innovative instrument designed as a two-dimensional interferometer for acquiring brightness temperature observations at L-band (1.4 GHz) globally and with a revisit time less than 3 days, a spatial resolution of smaller than 50km, and with a range viewing angles (0-50 degrees) for the estimation of soil moisture and ocean salinity, both are key variables used in weather, climate and extreme-event forecasting. As a secondary objective data acquired by SMOS over ice/snow regions may prove useful to characterise the ice and snow
layers and thus complement other satellite observations to advance the science of the cryosphere (see also
A number of open scientific questions related to the physics of the signal, perturbing effects and the retrieval concept by accounting on SMOS observational characteristics needed to be addressed to prepare for the mission. In addition, appropriate campaigns were designed and organised to provide suitable data for the analysis. This included the analysis of the physics of the signal (sensitivity towards ocean salinity, soil moisture and perturbing effects), the analysis of the data product requirements of the user community and the development of retrieval concepts. Campaign activities included the WISE, LOSAC, and EuroSTARRS campaigns.
This paper is intended to summarise the activities performed so far to advance our knowledge of the microwave radiation emitted by the Earth at L-band, the capability to retrieve soil moisture and ocean salinity from it and its dependence on other factors. It will also give an outlook of future planned activities to prepare for SMOS mission.
Determination and description of groundwater systems is essential for the management and development of ecological values, especially in the valley parts of river basins. At the land surface, groundwater systems appear as infiltration (relatively dry) and discharge zones (relatively wet). Groundwater discharge zones offer a high potential for nature values because of their constant moisture presence and their specific water quality. Current methods for the determination of discharge and infiltration zones use either detailed time-consuming fieldwork or data intensive numerical simulation models. Consequently, there is a direct need for repeatable, area covering, mapping possibilities for the determination of moisture gradients and more specifically discharge and infiltration zones. Within the framework of the CASI-SWIR measuring campaign 2002, the Department of Hydrology and Hydraulic Engineering of the Vrije Universiteit Brussel (VUB) executed an airborne hyperspectral remote sensing and field campaign in the Doode Bemde to analyze moisture gradients in the Doode Bemde, a riparian nature reserve. The main objective of the study is to test the best hyperspectral analysis method, using the hyperspectral CASI-SWIR data, for the known, based upon field and simulation data, moisture gradients in the Doode Bemde area. Simultaneously with the airborne hyperspectral campaign, field measurements of soil moisture, groundwater levels, vegetation temperature and spectral characteristics of some key vegetation species (phreatophytes) were performed. The method of analysis consists of statistical comparison of moisture gradients, obtained from measurements and simulations, with individual bands, a combination of bands and multivariate derivatives. The paper describes the set-up of the field and airborne measurement campaign, the methodology of analysis as well as first analysis results.
SPOT-Vegetation (SPOT-VGT) data at full spatial resolution were used in order to assess the potentiality and feasibility of using low resolution optical satellite data for the estimation of time/space dynamics of surface moisture content. The seasonal trend of diverse parameters (single channels or spectral indices) suitable/or
specifically designed for moisture estimation have been analyzed in some test sites of southern Italy. The investigations were performed by using imagery selected from a long time series of ten-day compositions of VGT data acquired from 1998 to 2002. The preliminary results indicate that the use of the VGT data for obtaining
information on the time/space dynamics of surface moisture content can be useful.
The Scanning Multichannel Microwave Radiometer (SMMR) provides ground coverage on the average of two times per week during the day and two times per week during the night. However, past problems associated with surface soil temperature estimation have resulted in both systematic and random differences between various applications utilizing both daytime and nighttime observations. This was especially true for SMMR-based soil moisture retrievals. Significant offsets were frequently observed, which prevented the daytime and nighttime soil moisture data from being combined into a single data set. This leaves one with a time series of data with very poor temporal resolution, and limits its usefulness for other applications. Improvements in surface temperature estimation have reduced the differences between day and night estimates significantly. The improved consistency between the two data sets now permits combining them into one, making the data more useful, especially for other land surface processes applications.
In the study, the NDVI and land surface temperature (Ts) were calculated using NOAA-AVHRR image to construct the NDV/-Ts space from which a dryness index-temperature/vegetation dryness index (TVDI) which was based on the slopes was suggested according the interpretation of NDVI-Ts SPACE. With the dryness index, the drought spatial patterns in China for every ten days in May 2000 were studied. The dryness index that combines the land surface temperature with vegetation spectral index is computationally straightforward because it was based on the information derived from satellite data only. Using the TVDI, the surface moisture status in May in 2000 was studied. The TVDI spatial pattern was compared with the measured topsoil moisture from the weather stations around China with the linear regression strategy. A negative linear correlation between TVDI and the measures soil moisture was found, thus TVDI’s validity in monitoring drought was verified. TVDI and CWSI that was based on Ts only were compared. Results showed that TVDI had a more remarkable relation than CWSI to soil moisture. So we can reach the point safely that TVDI based on the combinational information of Ts and NDVI was superior to CWSI solely based on Ts in monitoring regional drought.
A methodology was recently developed to estimate the land surface parameters soil moisture, soil temperature and vegetation optical depth on a global scale by using passive microwave remote sensing. This methodology is general, in a way that it does not require any field observations of soil moisture or canopy biophysical properties for calibration purposes, and can be used with microwave observations at different wavelengths. However, several algorithms in this approach are somewhat empirical, and the vegetation component in this methodology is still difficult to understand and interpret. A follow up field experiment was planned for April 2003 to address some of these issues. The experiment was conducted at a controlled meteorological field site in Wageningen (The Netherlands). Three different plots, a bare soil, a soil with short grass (reference site), and a site with growing grass vegetation were selected. Several hydro-meteorological parameters were monitored extensively at each site, including the radiobrightness temperatures from the ELBARA 1.4 GHz passive microwave radiometer. This paper gives a description of this field experiment and will demonstrate several effects of vegetation on the radiobrightness temperature.
Estimation of regional evapotranspiration is of major importance in hydrological modeling, where the partitioning of available energy into sensible and latent heat fluxes is crucial. Point-based measurements are routinely obtained with micrometeorological methods through a combination of radiometers and eddy-covariance instruments. Notwithstanding closure problems, they are considered to yield reliable flux point values. However, when dealing with heterogeneous semi-arid terrain, these point estimates are not representative for regional values.
In this paper the results are presented of an analysis where MODIS images are used for the mapping of energy and water balances of a heterogeneous land surface in a savannah environment on the southern fringe of the Okavango Delta (Maun, Botswana). Despite its semi-arid character, fresh floodwaters arrive through the Delta seasonally and therefore part of the area’s vegetation is always transpiring at a potential rate.
The model we implemented is governed by remotely sensed values of surface temperature, reflection and vegetation density. The availability of MODIS data provided an opportunity to test the new algorithm by determining the energy balance components on a regional scale for a heterogeneous area and then comparing the results with energy flux measurements using a meteorological flux tower situated in a woodland savannah environment.
The results indicate good estimates of net radiation, soil and turbulent fluxes. However, if energy closure problems are neglected, latent heat estimates show significant deviations.
The use of actual evapotranspiration derived by satellite data at watershed scale in water balance modelling of forested mountainous watersheds is studied. Mean monthly maximum composites of the Normalized Difference Vegetation Index (NDVI), derived from the National Oceanic and Atmospheric Administration’s (NOAA) / Advanced Very High Resolution Radiometer (AVHRR) were correlated with monthly actual evapotranspiration rates estimated by a water balance model. The water balance model was applied to three mountainous and forested watersheds of Central Thessaly in Greece and the actual basin-wide evapotranspiration was estimated using two methods for the estimation of basin wide precipitation and two methods of potential evapotranspiration. The derived values of actual evapotranspiration were then correlated to NDVI data, and the developed equations were validated temporally and spatially. The actual evapotranspiration estimates, derived from NDVI and used in the water balance model, resulted in equally accurate simulations of monthly runoff when compared with the simulations acquired from the classical application of water balance model.
This paper describes work towards building an integrated Earth sensing capability and focuses on the demonstration of a prototype in-situ sensorweb in remote operation in support of flood forecasting. A five-node sensorweb was deployed in the Roseau River Sub-Basin of the Red River Watershed in Manitoba, Canada in September 2002 and remained there throughout the flood season until the end of June 2003. The sensorweb operated autonomously, with soil moisture
measurements and standard meteorological parameters accessed remotely via land line and/or satellite from the Integrated Earth Sensing Workstation (IESW) at the Canada Centre for Remote Sensing (CCRS) in Ottawa. Independent soil moisture data were acquired from actual grab samples and field-portable sensors on the days of RADARSAT and
Envisat Synthetic Aperture Radar (SAR) data acquisitions. The in-situ data were used to help generate spatial soil moisture estimates from the remotely sensed SAR data for use in a hydrological model for flood forecasting.
The development of hyperspectral technologies in the infrared domain allows new methods of temperature emissivity separation to be elaborated. These methods, which are inappropriate for multispectral measurements, use the amount of information contained in a measured spectrum to increase the number of equations and stabilise the system to solve, or to differentiate the miscellaneous radiance components reaching the sensor thanks to the narrow bands of the spectrometer. Two techniques based on ground measurement are tested. These are the multi-temperature and the spectral smoothness methods. Both need very weak a priori hypotheses. The first one measures the radiance coming up from a surface for several temperatures (or at different times of the day) and solves the over-determinated system of equations. The only hypothesis is the invariance of emissivity between the measurements. The second one uses the difference of spectral variability between an atmospheric spectrum, which is composed of lines, and an emissivity spectrum, which is smoother. Both methods need the radiance at the surface level and the downward irradiance as inputs. Atmospheric corrections have to be made along the upward path between the surface and the sensor, so the atmosphere has to be accurately characterised (especially for the spectral smoothness).
This paper presents a numerical study of both methods (spectral smoothness and multi-temperature method). The effects of the radiative transfer on the retrieved emissivity are analysed and protocols to characterise the atmospheric contribution to the measured signal are described. Finally, a sensitivity analysis and an error budget of the methods are presented.
Land surface temperatures are important in global change studies, in estimating radiation budget, heat balance studies and as control for climate models. A new algorithm for estimating land surface temperature and emissivity spectra for multi spectral thermal infrared ranging from 8 to 12mm images has been developed recently (Schmugge et al., 2002) for use with data from the Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER) on the TERRA platform. Similar methods are also used with the MODIS instrument.
In this study, the method developed by Ogawa et al. (2002) was adopted to estimate the broadband emissivity from the narrow band emissivities of the five TIR channels of ASTER instrument in an area on the southern fringe of the Okavango Delta (Botswana). MODTRAN 4 was used to determine the necessary atmospheric corrections while software was developed to facilitate MODTRAN pre- and post-processing. The results were compared with field data, with a LANDSAT 7 image of the same day, and finally also with reported ASTER surface temperature and emissivities for the same image (high level ASTER product).
Results indicate that the surface temperature depends rather sensitively on atmospheric transmissivity. No relation was found between broad-band emissivity and NDVI, contrary, for example, to earlier findings in Botswana by Owe and Van de Griend (1993). Using the TES method it becomes possible to obtain more reliable solutions to the energy balance and evapotranspiration problem, especially in semi-arid areas.
This paper presents a framework for earth surface temperature retrievals from Multispectral thermal images acquired buy the US Dept. of Energy Multispectral Thermal Imager (MTI) satellite. The satellite has 15 spectral band including 5 bands in the thermal IR. Also there are three near IR bands for daytime retrieval of atmospheric water vapor to use for atmospheric compensation of the thermal IR measurements. Multispectral thermal IR techniques for retrieval of water surface temperature from MTI data have been implemented and made operational. The approach to temperature retrieval presented is amenable to land surface as well as water surface temperature retrievals. The approach uses the best available atmospheric profiles of temperature and humidity as inputs to MODTRAN4 to compute look up tables of the necessary atmospheric radiative transfer components. These profiles can come from any source but we make extensive use of grided data assimilation products for numerical weather prediction. For daytime images, the humidity profiles are scaled by the column water vapor retrieved from near IR bands on a pixel-by-pixel basis. Ground leaving radiance images can then be produced and apparent temperatures are computed given some source of emissivity information. To date, we have implemented the Normalized Emissivity Method (NEM) and the ASTER temperature emissivity separation approach adapted to MTI, and the ASTER and MODIS spectral libraries are available. A final retrieved temperature is determined from the temperatures retrieved for each band according to a specified rule. Validation of the techniques is presented for one water site and two land sites where the spectral emissivity is known and simultaneous ground measurements of skin temperatures were made. The daytime retrieval results are consistently within 1.1C of the ground measurements. Finally, we investigate the impact or using different global weather data products as the meteorological data source.
The objective of this study was to use satellite thermal data for the detection of areas with lower temperature in Alfios River Basin, Greece. Alfios is the biggest river of the Peloponnese. The climate of the area is typically Mediterranean with rainy winter and very hot summer.
Firstly we converted the radiation into Celsius Degrees in order to calculate the surface temperature. Then in order to better distinguish thermal deviations we have classified the temperatures in fifteen classes using the density slicing method.
We proceeded to the control of the model accuracy using in situ temperature measurements for the days of the images acquisition. The results showed that the error is less than 0.3 °C in the temperature range or about half a tone unit value of TM & ETM band 6 and it is acceptable.
We have mapped areas along the river channel that present 6-8 °C lower temperature than the rest of area. Even the third or second order branches of Alfios River create long lanes of lower temperature. Also the areas around the big artificial lakes of Pinios and Ladonas present significant lower temperature.
Then we have checked the effect of the burnt areas to the temperature. We have noticed that the burnt areas present 6-10°C higher temperatures than the neighborhood areas.
The general conclusion is that the TM & ETM thermal bands can be used for the creation of temperature profiles across large areas where the construction of climatological stations is difficult.
Two remote sensing techniques used to measure water vapor content in the atmosphere are presented: the Lidar/Dial technique and the GPS data analysis method. The dial method, as is well known, can be used to obtain range resolved measurements or an average concentration measurement on the long path using a target topographic method. This methodology permits measurement of the concentration of atmospheric trace gases and, in particular, water vapour profiles. The second remote sensing method is based on an application of the GPS (Global Positioning System). It enables the assessment of the signal propagation delay from satellites to ground-based receivers. Once ground temperature and atmospheric pressure are measured and the GPS signal delay is known, then an estimate of the columnar water vapour content can be performed. In this paper a comparison between the two remote sensing techniques of water vapour measurement are present.
Yield forecasts are of high interest to the malting and brewing industry in order to allow the most convenient purchasing policy of raw materials. Within this investigation, malting barley yield forecasts (Hordeum vulgare L.) were performed for typical growing regions in South-Western Germany. Multisensoral and multitemporal Remote Sensing data on one hand and ancillary meteorological, agrostatistical, topographical and pedological data on the other hand were used as input data for prediction models, which were based on an empirical-statistical modeling approach. Since spring barley production is depending on acreage and on the yield per area, classification is needed, which was performed by a supervised multitemporal classification algorithm, utilizing optical Remote Sensing data (LANDSAT TM/ETM+). Comparison between a pixel-based and an object-oriented classification algorithm was carried out. The basic version of the yield estimation model was conducted by means of linear correlation of Remote Sensing data (NOAA-AVHRR NDVI), CORINE land cover data and agrostatistical data. In an extended version meteorological data (temperature, precipitation, etc.) and soil data was incorporated. Both, basic and extended prediction systems, led to feasible results, depending on the selection of the time span for NDVI accumulation.
The problem of ecological agriculture production is a question of present interest all over the world. For the past few years in Bulgaria were performed a number of reforms associated with utilization and management of agricultural and forest areas. Contemporary the problem cover eco-monitoring, specification of the trends and rehabilitation criteria determination on the basis of satellite technologies.
The recent work presents a methodic for determination of suitable areas for sustainable ecoagriculture development. The main item in this methodic is test area definition based on remote sensing basis. It is proposed "point-polygon" method for correlation links between test point objects and certain test polygon. The test point objects are defined by GPS techniques. This approach enables complex quantitative assessment of the region with minimum resources consumption.
For precise results obtaining is performed interpretatation of multispectral and panchromatic aerophoto and satellite images of different time periods, it is specified crop phenology and field check. Data are integrated in GIS for more efficient monitoring of the characteristics of the region of interest.
The methodologies developed during these investigations can be applied to other regions, and have potential for providing modelers with extended data sets of independently derived agriculture data and monitoring studies.
Author(s): Victor I. Povkh; Eugeny A. Vorobeychik; Ludmila A. Shljakhova; Gennady P. Garbuzov; Irina E. Beljaeva; Inessa V. Sytnyck
The opportunity of regular receiving of the MODIS/TERRA radiometer data in the direct broadcast mode makes it possible to use these data as perspective information source for the detailed analysis of agricultural production on a regional scale. The most suitable MODIS/TERRA data subset for this purpose is a pair of red and NIR spectral bands with 250 m spatial resolution. The specified spectral ranges are in accordance with the well known remote sensing
algorithms applied to estimate vegetation properties. In the SRIA-Centre, Rostov-on-Don supported by the Rostov region administration, techniques have been tested for several years aimed at prompt checking of winter wheat crops by the MODIS/TERRA radiometer data. The type of crops selected for the related studies is due to the large size of individual fields of winter wheat in the region. Thus, any losses of accuracy resulted from the low spatial resolution of
MODIS/TERRA radiometer data are diminished. To achieve the needed accuracy, the quality assessment of land vector layers is of great importance. These vector layers are used both for registration of space images and for field masks creation during image interpretation. The vector layers have been created on the basis of the Russian high resolution space photo of the KFA-1000 starting from 1999. Thematic analysis of MODIS/TERRA radiometer data is made with use of supervised classification techniques which require to define training classes prior to execution. Training classes are determined using the selected test sites (crop fields). The numerous regular inspected test sites are sorted in various agricultural zones of the region. GPS receivers serve to make analyses of the inspected fields and probe points location. The scaling test of winter wheat crops monitoring technology was carried out in 2002. Comparison of its
results and farm reports has shown a sufficient reliability of remote sensing estimations on both sizes of the areas of agricultural production and crop condition assessment. All remote sensing processing procedures are conducted by ENVI 3.5 software.
In this paper a study carried out by airborne hyperspectral data is presented. Image data concern a forest area (60 km North of Rome -Italy), to evaluate how different spatial resolutions can affect vegetation spectral response and therefore the discrimination among the different communities.
MIVIS images were acquired at different flight altitudes (2000 m and 5000 m), in the same day and on the same surface targets. A radiometric field survey was carried out in order to radiometrically calibrate the airborne images.
Classification were performed on each image by two different techniques: Maximum Likelihood and Spectral Angle Mapper. The results of these classification methods were analysed to evaluate how different spatial resolution can affect vegetation spectral response.
In particular the relationship between spectral and spatial resolution of hyperspectral images was investigated by resampling the 2000 m image (4m/pxl) in order to simulate the radiometric response of surface targets in the MIVIS image acquired at 5000 m of altitude (10m/pxl).
The result suggest that the spatial resolution of aerial images must be decided when the overflight is planned, because the resampled data keeps the original spectral characteristics and upscaling methods do not provide meaningful data in heterogeneous area.
This paper focuses on application of artificial neural networks (ANN) in land suitability evaluation. There are some problems in applying fuzzy system to land suitability evaluation such as self-adjusting ability of the membership functions and rules of fuzzy evaluation system. In this paper, the model of fuzzy neural network is designed for land suitability evaluation. This model is the result of integrated fuzzy system and artificial neural network. This fuzzy neural network model has five layers. The learning algorithm of the model has been designed based on the principle of error back propagation of neural networks. The learning strategy, algorithm and efficiency of the model have been tested and the results of test are satisfied.
This paper studies landuse change model based on cellar automata of decision-making with grey situation. the urban land use dynamic change model is constructed based on decision-making with gray situation and hierarchical method as well as CA, in which the hierarchical method is used for dealing with the importance of among objects and factors. The arable land is used as constraint factor for transformation rules in the model, Lastly, the model has been tested by taking Qionghai city of Hainan province as an example, the result shows that the model is valid for simulating urban landuse change.
The values of the Normalized Difference Vegetation Index obtained from NOAA Advanced Very High Resolution Radiometer (AVHRR) have often been used for forestry application, including the assessment of fire risk. Forest fire risk estimates were based mainly on the decrease of NDVI values during the summer in areas subject to summer drought. However, the inter-annual variability of the vegetation response has never been extensively taken into account. The present work was based on the assumption that Mediterranean vegetation is adapted to summer drought and one possible estimator of the vegetation stress was the inter-annual variability of the vegetation status, as reflected by NDVI values. This article presents a novel methodology for the assessment of fire risk based on the comparison of the current NDVI values, on a given area, with the historical values along a time series of 13 years. The first part of the study is focused on the characterization of the Minimum and Maximum long term daily images. The second part is centered on the best method to compare the long term Maximum and Minimum with the current NDVI. A statistical index, Dynamic Relative Greenness, DRG, was tested on as a novel potential fire risk indicator.
The European Space Agency's Soil Moisture and Ocean Salinity (SMOS) Earth Explorer Opportunity Mission will be launched in 2007. Its goal is the global and frequent measurement of soil moisture over the land and surface salinity over the sea, two key parameters governing the complex global climate. SMOS’ single payload is the Microwave Imaging Radiometer by Aperture Synthesis (MIRAS), the first space-borne interferometric radiometer. SMOS will provide brightness temperature data over a wide range of incidence angles at vertical and horizontal polarizations (dual-polarimetric mode) or the full Stokes emission vector (full-polarimetric mode), from which the geophysical parameters will be derived. This paper focuses on the soil moisture retrieval problem using dual or full-polarimetric information. In this case, the brightness temperatures, as measured by the radiometer, depend mainly on five parameters descriptive of the surface under study: vegetation opacity and albedo, and soil surface temperature, roughness and moisture. Some of these parameters can be derived from other sensors or can be inferred from the multi-angular brightness temperatures themselves. Simulation results using the SMOS End-to-end Performance Simulator (SEPS) developed at the Universitat Politecnica de Catalunya (UPC) will be presented and discussed.
The CERES instruments obtain their measurements by three scanning thermistor bolometers: total channel which measures radiation in the 0.3 - 100 μm spectral region; window channel which measures radiations from 8 - 12 μm; and the shortwave channel which measures the reflected solar energy from 0.3 - 5.0 μm. The Aqua dedicated to advancing our understanding of Earth's water cycle and our environment was launched on May 4, 2002. The Aqua CERES Internal Calibration module follows similar operations as previous CERES instruments (FM1 and FM2) that were launched on the Terra spacecraft in December 1999. The Internal Calibration mechanism and the resulting data for Flight Model 3 (FM3) and Flight Model 4 (FM4) instruments are critical in determining the stability of both instruments as referenced to ground-based calibration. In this paper, the Internal Calibration procedure and the results will be presented. Both the FM3 and FM4 instruments were found stable within 0.3% since the launch date. Results for the FM1 and FM2 instruments will also be discussed.
Correlation between weather radar reflectivity and precipitation data collected by rain gauges allows obtaining empirical formula that can be used to create continuous surface of the rainfall. This surface can be used in distributed hydrologic modeling and early warning system in flood management. In this study rain data from multiple weather stations were correlated with reflectivity values from the radar covering the area for each time interval of the selected
rain event. Internet sources provided real-time precipitation data and images of weather radars for the continental United States, collected by United States Geological Survey (USGS) and National Oceanographic and Aeronautic Agency (NOAA). Database of 82 radar stations and more than 1500 rain gauges for the Continental part of USA was compiled and used for continuous downloading of radar images and rain data. Image sequences corresponding to rain events were extracted for two randomly selected radar stations in South and North Carolina. Rainfall data from multiple gauges under the radar zone (120 miles) were extracted and combined with corresponding reflectivity values for each time interval of the selected rain event. Results of regression analysis showed significant correlation between rain gauge data and radar reflectivity values and allowed derivation of empirical formula.
WATERMED project contributes to the international efforts in analyzing efficiency in water use, in particular for the Mediterranean Basin countries. The general aim of this project is to develop a comprehensive method for the study of the water use and the resistance to the drought of the natural and irrigated vegetation in the Mediterranean Basin, by means of a combined historical remote sensing database, vegetation models and field measurements. The project has provided regional maps of critical parameters at regional scale, such as land surface temperature, emissivity, and NDVI. A multi-temporal analysis using the PATHFINDER AVHRR land data has been carried out into the frame of this project to map and monitor changes in the biophysical characteristics of land cover over the last 20 years. On the other hand, REANALYSIS data, which is a result of a joint project NCEP/NCAR, have been incorporated to provide mean monthly climate data over the study area.
Author(s): Joaquim Fortuny-Guasch; Alberto Martinez-Vazquez; Juan M. Lopez-Sanchez; J. David Ballester-Berman
A coherent electromagnetic model developed for estimating the radar backscatter from rice crops is presented. This model is based on the first order solution of the polarimetric backscatter response as a function of the sensor parameters and the physical description of the rice plants. This paper presents a comparison
between simulations obtained by the model and experimental data collected at the European Microwave Signature Laboratory (EMSL). This laboratory is currently carrying out several campaigns of measurements over different growing stages of a rice crop sample. For the first stage of growing, simulated and measured backscattering coefficients, at HH, HV and VV channels, show a reasonable agreement. The sensitivity and dependence of the radar signal with respect to the input parameters (incidence angle, frequency and morphological characteristics of the rice plants) has been studied in order to achieve a future parameter inversion. The final aim of this work is to establish a reliable inverse algorithm to retrieve biophysical parameters of rice crops from radar measurements.
The strict link between intra-annual vegetation dynamics (phenology) and Earth's climate makes phenological information fundamental to improve understanding and models of inter-annual variability in terrestrial carbon exchange and climate-biosphere interactions. In order to monitor phenology in a landscape characterized by heterogeneous features rapidly changing over the territory, we performed multitemporal classifications of NDVI-AVHRR data and interfaced them with Landsat-TM data and orography. The sample area is the Vulture basin (Southern Italy), where cultivated and densely vegetated areas coexist with urban and recently built industrial areas. These land cover patterns rapidly change over the territory at very small spatial scales; it is a complex zone very interesting for studying the use of remote sensing techniques in the integrated monitoring context. Clusters having homogeneous NDVI time behaviors were identified. In spite of its spatial resolution, AVHRR NDVI effectively picks up the characteristic phenology for different covers and altitudes. Moreover, some pixels having particular microclimate were clustered and their characterization was only possible by using orography and TM classification information. The comparison of two intra-annual classifications (1996 and 1998) showed that the proposed approach can be very useful for studying change in pattern of vegetation dynamics.
Great attention is now paid to ecology of the environment, in whic plants are of great importance. However the present methods of biophysical analysis of plant states are very labor-intensive and require a lot of time. The structure of protein-pigment complexes is known to break in different dissolvents that results in the shift of maxima of chlorophyll absorption and fluorescence bands. That is why development of methods for remote diagnostics of plants is of great scientific and practical interest. They would make it possible to determine species and state of plants rather quickly and accurately. We have developed a setup and methods for optical diagnostics of the physiological state of plants to investigate the dynamics of the fastest part of fluorescence of plants in vivo. The method of laser-induced fluorescence makes it possible to observe the level of vegetative development of living plants, as well as their state under the impact of some stress factors.
In this paper, it is shown the importance of thermal measurements to characterize different surfaces carried out in boreal environment in the SIFLEX (Solar Induced and Fluorescence Experiment) campaign. The data was acquired in Sodankyla (Finland), over the boreal forest, from 23rd April to 10th June 2002. Bio-geophysical parameters such as land surface temperature and emissivity were retrieved in relationship with other parameters from fluorescence measurements made with a Passive Multiwavelength Fluorescence Detector (PMFD).
The thermal measurements of different targets (soil, vegetation, sky) under different observation angles have been carried out using a four-band field radiometer (CIMEL CE312) and two single band radiometers (EVEREST 3000.4ZLC and RAYTEK ST6). Angular measurements and transects have been also carried out concurrently to the satellites flights over the region.
Fragmentation indices derived from remotely sensed data are being increasingly used for landscape condition assessment and land cover change characterization. However, it is not yet fully understood how fragmentation indices are affected by spatial resolution, and the lack of comparability across scales seriously limits the potential usefulness of this kind of quantitative analysis of landscape patterns. We here consider a wide set of commonly used fragmentation indices and analyze the ability of aggregation filters for replicating the pattern and indices values corresponding to coarser spatial resolution sensors. We analyze simultaneously gathered Landsat-TM and IRS-WiFS satellite images, as well as TM patterns aggregated to coarser resolutions through standard mean and majority filters and through filters that incorporate the specific point spread function of the WiFS sensor. All the images were classified in forested areas, agricultural lands and water bodies for the computation of the fragmentation indices. We show that mean and majority filters tend to produce clearly more fragmented patterns than actual sensor ones. We found that incorporating point spread function in the aggregation process allowed to considerably improve the comparability of fragmentation estimations across spatial resolutions. The biggest improvement was found for indices like number of patches, edge length and mean patch size, which are the most sensitive to changes in spatial resolution and minimum mapping unit. On the contrary, indices like largest patch index, patch cohesion or landscape division were little affected by spatial resolution and did not show significant differences between the aggregation filters considered. Higher aggregation errors were found for water bodies than for forested areas or agricultural lands.Fragmentation indices derived from remotely sensed data are being increasingly used for landscape condition assessment and land cover change characterization. However, it is not yet fully understood how fragmentation indices are affected by spatial resolution, and the lack of comparability across scales seriously limits the potential usefulness of this kind of quantitative analysis of landscape patterns. We here consider a wide set of commonly used fragmentation indices and analyze the ability of aggregation filters for replicating the pattern and indices values corresponding to coarser spatial resolution sensors. We analyze simultaneously gathered Landsat-TM and IRS-WiFS satellite images, as well as TM patterns aggregated to coarser resolutions through standard mean and majority filters and through filters that incorporate the specific point spread function of the WiFS sensor. All the images were classified in forested areas, agricultural lands and water bodies for the computation of the fragmentation indices. We show that mean and majority filters tend to produce clearly more fragmented patterns than actual sensor ones. We found that incorporating point spread function in the aggregation process allowed to considerably improve the comparability of fragmentation estimations across spatial resolutions. The biggest improvement was found for indices like number of patches, edge length and mean patch size, which are the most sensitive to changes in spatial resolution and minimum mapping unit. On the contrary, indices like largest patch index, patch cohesion or landscape division were little affected by spatial resolution and did not show significant differences between the aggregation filters considered. Higher aggregation errors were found for water bodies than for forested areas or agricultural lands.
The present paper presents a method to characterize typical crop rotations from temporal series analysis of land use maps derived from supervised classifications of Landsat TM images. The analysis is based on spatial cross-tabulation of land use maps in raster format. As a case study, a temporal land use map series from 1993 to 2000 of the Flumen irrigation area (Huesca, Spain) was considered. The spatial cross-tabulation analysis between each pair of consecutive land use maps, performed in Idrisi 32, yielded a two dimensional matrix that allowed the identification of the typical crop rotations in the study area. Those are rice - fallow land - rice, sunflower - winter cereals - alfalfa - corn, and others as winter cereal or sunflower - fallow land - corn or alfalfa. Rice appears as a typical crop in this area, in which it is usually associated to salt- and/or sodium-affected soils. Those typical rotations have been also spatially located and represented in a map following the crop changes from one year to another year that are registered in the cross-tabulation images. The method can be useful to identify tendencies in the temporal variation of crop rotations in agricultural areas, and to locate typical areas with salt- and/or sodium-affected soils by mapping rotations in which rice is the main crop.
A new remote sensing application was developed and incorporated to the Agrarian Integrated Information System (S.I.I.A), project which is involved on integrating the regional farming databases from a geographical point of view, adding new values and uses to the original information. The project is supported by the Studies and Statistical Service, Regional Government Ministry of Agriculture and Fisheries (CAP).
The process integrates NDVI values from daily NOAA-AVHRR and monthly IRS-WIFS images, and crop classes location maps. Agrarian local information and meteorological information is being included in the working process to produce a synergistic effect.
An updated crop-growing evaluation state is obtained by 10-days periods, crop class, sensor type (including data fusion) and administrative geographical borders. Last ten years crop database (1992-2002) has been organized according to these variables.
Crop class database can be accessed by an application which helps users on the crop statistical analysis. Multi-temporal and multi-geographical comparative analysis can be done by the user, not only for a year but also for a historical point of view. Moreover, real time crop anomalies can be detected and analyzed. Most of the output products will be available on Internet in the near future by a on-line application.
Multiangular and hyperspectral capabilities of the last generation of remote sensing sensors require new data processing algorithms that can take advantage of this new type of information. In terms of atmospheric correction, taking into account surface directional reflectance properties leads to a coupling between surface and atmospheric radiative transfer effects that cannot be analitically decoupled in the most general case, so other strategies must be
developed. In addition to this, commonly used radiative transfer codes are based on a plane-parallel atmosphere approximation, what causes problems for large view and illumination zenith angles. The aim of this paper is to present an atmospheric correction method based on Vermote et al. scheme for MODIS atmospheric correction.
It considers BRDF effects in the surface, improving 6S code calculations in off-nadir configurations . We have simulated the top-of-the-atmosphere reflectances using nine different natural surfaces by means of MODTRAN4 radiative transfer code. The reflectance angular pattern retrieved for each surface has allowed us to validate the
model and check the improvements versus the original MODIS algorithm.
The objective of this work was to study the influence of the direction of the corn sowing in its reflectance and in the Green Normalized Difference Vegetation Index - GNDVI, seeking to supply necessary information to make possible a system of located application of nitrogen (N) in real time. The corn was sowed being used the technique of the direct planting, being administered the nutrients in agreement with the soil need and with a constant N rate of 160 kg/ha. The sowing lines were located in the North-South (NS)
and East-West (EW) directions. It can be concluded that the corn sowed at EW presented smaller reflectance and GNDVI and they were less dependent of the time than the corn sowed at NS. The GNVI showed more interesting appropriate for use in real-time variable-rate application system for N fertilizer based on canopy reflectance for showing less sensitive the sowing direction.
It is known that leaf reflectance spectra can be used to estimate the contents of chemical components in vegetation. Recent novel applications include the detection of harmful biological agents that can originate from agricultural bioterrorism attacks. Such attacks have been identified as a major threat to the United States’ agriculture. Nevertheless, the usefulness of such approach is currently limited by distorting factors, in particular soil reflectance.
The quantitative analysis of the spectral curves from the reflection of plant leaves may be the basis for the development of new methods for interpreting the data obtained by the remote measurement of plants. We consider the problem of characterizing the chemical composition from noisy spectral data using an experimental optical method.
Using our experience in signal processing and optimization of complex systems we propose a new mathematical model for sensing of chemical components in vegetation. Estimates are defined as minimizers of penalized cost functionals with sequential quadratic programming (SQR) methods. A deviation measure used in risk analysis is also considered.
This framework is demonstrated for different agricultural plants using adaptive filtration, principal components analysis, and optimization techniques for classification of spectral curves of chemical components. Various estimation problems will be considered to illustrate the computational aspects of the proposed method.