Societies in general are becoming keenly aware of the need to manage information from a geographic perspective. This awareness has been brought about by the twentieth century trends toward a global community and economy. Throughout the world, governments, utilities, and businesses are investing billions of dollars in computer systems that store, manage, and analyze maps and geographic information. Geospatial Information Technology (GIT) has grown in tandem with the widespread development and use of increasingly inexpensive, powerful computers.

Far sighted people throughout the world are beginning to apply GIT in attacking the afflictions of crime, poverty, disease, pollution, urban congestion and last but not least for agriculture.

This technology has not been fully implemented in Malaysian agriculture but has been widely used today in land surveying and in forest resource and management such as recreation, land cover mapping, rehabilitation, land use, deforestation, peat swamp forest changes and detecting forest disturbance and cover type (Kamaruzaman, 1999; Kamaruzaman and Aswati, 1999; Kamaruzaman and Zulhazman, 1997). At the frontier of the information age, GIT is emerging as a powerful means to manage voluminous geographic data, to help cope with the information explosive, and to provide a foundation for solving the problems that beset planet Earth and its inhabitants.

The objective of this paper is therefore to highlight the potential of GIT for Malaysian agriculture in the next millenium, particularly the use of remote sensing, Geographic Information System (GIS) and Global Positioning System (GPS) for site-specific agriculture management.
 
 

Site-specific management consists of three basic elements: global positioning system (GPS), remote sensing (RS) and geographic information systems (GIS).
 

1. Global Positioning System For Agriculture

 

 

The Global Positioning System (GPS) is a highly accurate satellite based radio navigation system providing three-dimensional positioning, velocity, and time information. GPS is an all-weather system that provides continuous and worldwide coverage.

GPS is based on a constellation of 24 satellites orbiting the earth at very high altitude (Figure 1). In a way, the satellites so-called "man-made stars" are to replace the stars that we have traditionally used for navigation (Figure 2). The satellites are high enough that they can avoid the problems encountered by land based systems and they use technology accurate enough to give pinpoint positions anywhere in the world, 24 hours a day (Jeff, 1989).
 
 




Figure 1: GPS nominal constellation satellites
 


Figure 2: GPS satellite

GPS receivers use signals from four or more satellites (Figure 3) in view to display or save the users or farmers position, velocity, time, and other data as needed for their precision agricultural applications. This data is collected on memory cards along with yield data for use in the office by a GIS software package. Other agricultural applications of GPS include controlling the application of agricultural chemicals, locating insect or weed problems, field preparation, and controlling planters.


Figure 3: GPS receivers on board vehicles use signals from four or more satellites





Differential GPS (DGPS) is used for precision agricultural mapping applications to overcome the effects of selective availability (SA) and atmospheric signal distortions (Figure 4). Depending on the technique and the kind of GPS receivers used, DGPS can provide a position accuracy of about 10 m to better than 1 cm, and speed accuracy of better than 0.0001 km/hr. DGPS corrections are available for both real-time and post-processed applications from a variety of public and commercial sources which will meet the requirements of many agricultural mapping users.


Figure 4: DGPS signals to overcome the effect of SA and atmospheric disturbance.





A positioning system is essential to the implementation of site-specific farming. Essentially, the GPS required a position resolution in the range 0.2 m up to 20 m (Saunders et. al., 1996), reliably maintained in all areas of the field. The GPS available in UPM is a hand-held Trimble’s GeoExplorer™ II (Figure 5). It is a pocket sized GPS receiver with user-friendly processing software designed for mapping and GIS data collection with positions accurate from 2 to 5 m after differentiation.

Trimble has a number of different GPS receivers that may be useful in precision agricultural applications. These models are the GeoExplorer II, ASPEN XL, Direct GPS, and Pro XL GIS data capture products, the GPS Pathfinder real-time Community Base Station, the TrimFlight System, the AgGPS 122 combined with GPS/MSK beacon receiver, DSM, and 7400Msi GPS sensor models.


Figure 5: Hand-held Trimble’s GeoExplorer™ [A] and GPS Pathfinder™ Pro XR [B].




2. Remote Sensing For Agriculture

 

 

Remote sensing is an observation of an object without actually being in physical contact with it (Barett and Curtis, 1976; Maguire, 1989). It consists of a measurement and recording of an electromagnetic energy, which is bound back from the earth surface and atmosphere using a sensor attached in an aeroplane or satellite. Detail identification from the earth and atmosphere will be made based on the measurement or observation gathered from the sensor. Usually data in the form of digital used for producing an image. Those digital data were analysed more efficiently (display, enhance and manipulate) using a computer.

Almost all of the applications of remote sensing to date have been based on observing crops in distinct areas of the electromagnetic spectrum (Figure 6). Agricultural remote sensing is commonly done in the visible, near-infrared and thermal infrared portions of the spectrum; however, new applications in the microwave area are under development.
 
 




Figure 6: Electromagnetic spectrum

An instrument called spectroradiometer measures the amount of energy radiating from a surface in a particular portion of the spectrum. Spectroradiometers can be hand held for research purposes and monitoring small field plots or places on board aircraft and satellites to survey entire fields, farms, or agricultural regions (Figure 7a and 7b). The airborne systems can image large areas quickly, safely, and cost-effectively, providing a more complete picture than can be obtained from outer space or aerial photography. The amount of radiation from an object (called radiance) is influenced by both the properties of the object and the radiation hitting the object (irradiance). When the sun is used as the source of irradiance, the irradiance is not constant (varying with the time of day and atmospheric conditions); therefore, the radiance of an object is not necessarily a good indicator of physical properties. Instead, the apparent reflectance of an object is best used to learn about its properties. Reflectance is the ratio of the radiance from an object to the irradiance reaching the object. The reflectance of an object can be impacted by the angle we are looking from and the angle of the sun. The radiometer has special filters that can be designed so that only radiation from a specific part of the spectrum is measured.
 
 





Figure 7a: Spectroradiometer on-board aircraft [A] for scanning technology and data analysis methods to detect stressed vegetation, monitoring crops and mineral deposits from the air.
 
 





Figure 7b: Hand-held spectroradiometer with simple menu-driven programs control the set-up, acquisition, and data manipulation functions and easy-transfer to other software.
 

The usefulness of the visible and infrared portions of the spectrum can be seen by comparing the typical spectral responses of a crop canopy and a bare soil (Figure 8).

Notice the peak in crop curve around 550 nm, this corresponds to the color green. That is why plants look green to our eye, as this is where they reflect the most radiation in the visible part of the spectrum (Figure 8). Secondly, notice the dip in the crop spectra around 690 nm, where this corresponds to the color red and is primarily due to chlorophyll absorption. The crop canopy is lower in the red compared to the soil but much higher in the near-infrared (NIR). The structure of the leaves accounts for this relative increase in reflectance in the NIR. Because of these properties, measurements of reflectance in the red and NIR can be used to determine differences in crop canopy densities. The response of vegetation in the red and NIR have been used to derive "Vegetation Indices", which typically involve some ratio of NIR to red reflectance.

Another area of the spectrum that is useful for assessing crop conditions is the thermal portion of the spectrum. Measures of radiance in this area can be used to derive the surface temperature of the crop. Plants take up water from the soil and then release it from their leaves back to the air. As the water transpires from the plant, its leaves are cooled (the same way we are cooled by our perspiration). If a crop cannot get enough water, its surface temperature will increase, so the plants surface temperature can tell us something about how healthy it is.

Soil physical properties such as organic matter have been correlated to specific spectral responses (Dalal and Henry, 1996; Shonk et al., 1991). Leone et al. (1995) reported that multispectral images have shown potential for the automated classification of soil mapping units. Figure 9 showed a gray-scale image in the red portion of the spectrum. Percent sand and clay in the top 30 cm of the soil horizon is displayed over the approximate location of point samples taken by Post et al. (1988). Note that the brighter portions of the image correspond to areas of high sand content.



Figure 9: Gray-scale image of a fallow field in the red portion of the spectrum with point measurements of percent sand and clay shown over the approximate sampling locations.
 
 

3. Geographic Information System For Agriculture

1. Reduce Labor

 

 

The expectant/desire to reduce labor and to maintain or increase agriculture production is one of the factors forming a strong influence in using GIT in agriculture. For example, in order to get all the water level or data from a large agriculture area, formerly the whole area will have to be manually ground surveyed (Figure 16a). This will increase labor or manpower. Now, with the capability of remote sensing, as one of GIT components, the whole area will be classified (Figure 16b). With that, only several classified areas will be sampled. By reducing manpower, labor cost also reduces with higher time consuming.
 
 





Beside labor for data collection, labor for farm operations also will be reduce because with all the information gathered spatially and precisely using remote sensing, GIS and GPS, farm operations such as pest and disease control management, weed management, irrigation management etc will focused only at certain problem spots in the farm.
 
 

2. Higher Environmental Constraints

 

 

The need for a sustainability agriculture environment has also lead today's society to use GIT. For example, the 16-day orbital cycle for LANDSAT has allowed user to have continuing satellite image for their regular farm/land observation. This definitely will prevent land degradation and land destruction due to no ground survey or ground checking being performed. Less degradation and destruction will reduce erosion and compaction of soil and finally agriculture area will be environmentally sustained.
 
 

3. Higher Need For Information

 

 

Farmer needs information from down below the soil such as nutrients variability, water content, porosity, pest and disease existence to high above the soil such as climatic changes. This is also one of the factors that affect further use of GIT in agriculture. The shape of reflectance spectrum can be used for identification of plant health condition and vegetation type. For example, the reflectance spectra of plant 1 and 2 in Figure 17 can be distinguished although they exhibit the generally characteristics of high NIR but low visible reflectance.



Plant 1 has higher reflectance in the visible region (0.4-0.7 µm) but lower reflectance in the near infrared (NIR) region (0.7-0.9 µm). For the same plant, the reflectance spectrum also depends on other factors such as the leaf moisture content and health of the plants. These properties enable the whole plant/vegetation condition to be monitored using satellite remote sensing images. As a result, the information of unhealthy plant can be detected. Healthy plant (high content of chlorophyll in the leaf) will reflect red to near-infrared (NIR) EM spectrum (0.7-0.9 m m wavelength) in remote sensed image/data. While unhealthy plant will not reflect reddish color instead it reflects more towards bluish color indicating the low chlorophyll content of the leaf. Uninfected plant but aged and young plants also have limited content of chlorophyll. Due to both aged plant and young plant have same reflection of EM spectrum with stressed and unhealthy plants, immediate ground check has to be conducted.
 
 

4. Improvement In Computerization

 

 

Information vital for agriculture operations or agriculture maintenance can be obtained from manual, maps and experiences. Now, with the convenient assessment-using computer, all the information includes manual, maps, experience and spatially gathered information such as satellite or airborne data can be put together in one computer personal unit. These informations can further be analyzed and manipulated to gain new information such as yield estimation.
 
 

5. Lower Development Cost And Higher Agriculture Production

Although high investment in capital especially in getting hardware and software, presumption in having lower development cost but higher agriculture production has lead user to further use GIT in agriculture. Development costs are lowered with early detection of disease or pest or climatic problem occur and with that early prevention from spreading. Development cost are minimized such as cost for weeding, disease/pest control and other cultural management due to the early and accurate detection. As a result, higher agriculture production will be obtained from immediate prevention and plant health management.

AGRICULTURE

Under the Third National Agricultural Policy (NAP3) that covers the period 1998 to 2010, marks another milestone in strategizing agriculture as an important sector in the economic development for Malaysia. The agriculture sector will be revitalized as a sector, which not only supplies food for the population but also raw materials for downstream manufacturing industries. In the suit for agriculture excellence, the agriculture sector has to adopt new technologies such as remote sensing, GIS and GPS. In addition, the agricultural land use has increased from 5.0 million ha in 1985 to 5.8 million ha in 1995 which, has indirectly raised new issues and challenges for Malaysian agricultural managers to manage such development. This has led them to seek new developed technologies such as GIT.

Weed mapping is another area of concern to utilize GIT in the next millenium. In conventional farming, each field has been considered to have a uniform weed distribution and herbicides have been applied at a uniform rate. Patch spraying recognizes this variation by applying herbicides according to locally determined requirements. The advantages of patch spraying are optimal use of herbicides and reduction of herbicide costs and environmental impact. Weed maps can be estimated by co-kriging when weed densities from previous years and soil factors as a secondary variable are used. Despite reduced sampling may result in poorer estimated weed maps, co-kriging based on reduced sampling and using either soil maps or old weed maps could be a shortcut to make better estimates of weed maps.

Another important potential GIT utilization in agriculture for the next millenium is the mapping of soil nutrient variability in oil palm, cocoa and rubber plantations. The objective is basically to gather the soil nutrients variability and nutrient content in oil palm, cocoa and rubber tissues The soil nutrient content and the nutrient content of these crops can then be integrated into a GIS, with some correlation models and variogram to produce base map of the plantation. By correlating the data gathered from remote sensing and GIS/GPS techniques, the crop yield of the study area can be predicted using yield maps analyzed using geostatistical methods. The semivariogram analysis was proven to be an efficient tool to determine the spatial and temporal structure of yield data. The potential yield can be represented by soil survey maps, so this data can be used to determine site-specific nitrogen rates, as a cost effective alternative to the time consuming and cost intensive N_min determination. Even though it will take sometime to map yield, the results of the study should show the potential of yield maps to provide site-specific information about yield potential, soil characteristics and other factors influencing the yield pattern. For site-specific information on soil nutrients (P, K, Mg) and pH, a one hectare soil grid sampling proved to be a cost effective method for larger fields (>5 ha) (Lütticken, 1997).

Another potential GIT application in Malaysian agriculture is the mapping of root zone capacity by co-kriging aerial photographs and point measurements of available soil moisture. The root zone capacities at approximately 20 grid points in each field can be determined. The accuracy of prediction by co-kriging was not improved by including soil texture when compared to ordinary kriging or using the inverse distance method. In years when the root zone capacity had a large influence on plant growth due to water stress, the prediction of root zone capacity by co-kriging point measurements of root zone capacity to reflection measured by 'false red' aerial photographs should work reasonably well for all kinds of agriculture crops.

Another potential GIT application in Malaysian agriculture is the mapping of Malaysian Agriculture Green Fields. This application is to map the entire major plantation crop throughout Malaysia using satellite imagery and airborne remote sensing. For instance, SPOT 4's high-resolution vegetation sensor has the capability to monitor crops continuously, holding much promises for all kinds of application at a variety of scales. The multispectral SPOT imagery with reasonable high spatial resolution (20m) combines with GPS technology can produce detailed mapping of field parcels and accurate measurement of areas under cultivation. Initial crops yield estimates can be made three to four months before harvesting because variations in multispectral response throughout the crop growth cycle enable cropping patterns to be identified and classified.
 
 
 
 

As quoted by our Prime Minister of Malaysia in NAP3, "Malaysia is currently the world leader in research of tropical agricultural products". It is therefore vital to continue revitalizing our agricultural industry by the operationalization of GIT at the farm and state management level. By implementing such technology, we hope to have safe, nutritious and high quality food at affordable prices. In addition, future Malaysian agricultural sector is expected to reduce labor requirements, maximize land utilization, highly competitive, enhance private sector investment in food production and transform the smallholders into a more commercial sector towards sustainable agricultural development. Vegetation data gathered from satellite imagery can be geared to many potential GIT applications such as monitoring and forecasting of major crop yields; crop status reporting; studying the impact of drought, flooding and disease, and detecting crop anomalies; studying agro-climatic conditions, which could aid to a wise and sensed agricultural decision-making.