| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Texas A&M University Agricultural Research and Extension Center, 10345 Agnes St., Corpus Christi, TX 78406
* Corresponding author (cj-fernandez{at}tamu.edu)
Received for publication June 30, 2005.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONSTRUCTURAL AND DEVELOPMENTAL...IMPLEMENTATION OF A WEB-BASED...CONCLUSIONSREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONSTRUCTURAL AND DEVELOPMENTAL...IMPLEMENTATION OF A WEB-BASED...CONCLUSIONSREFERENCES |
|---|
Since the advent of computing systems, and particularly since personal computers became widely available in the early 1980s, numerous computer-assisted numerical simulation models were developed to quantify the environment and the development and growth of crops (e.g., Whisler et al., 1986; Hanks and Ritchie, 1991; McKinion et al., 1995; Cheeroo-Nayamuth, 1999). Computer-assisted simulation models, in particular, mechanistic-type simulation models, can be effective tools to integrate current knowledge or ideas about interactions between crops and the environment (Loomis et al., 1979; Loomis, 1985), and several of these models have become useful tools for research and education. Only a few of these models, however, have been developed as decision support systems for farming (e.g., Lemmon, 1986; McKinion et al., 1989; Hearn, 1994; Cochran et al., 1995; Landivar et al., 1995; Hearn and Bange, 2002). Computer-assisted decision support systems can be important tools in the decision-making process in farming (Ritchie, 1995), as they can assist crop managers deal with environmental variability and the complex nature of soil–crop–pest–environment interactions.
The task of transferring technology to farmers using computer-assisted numerical models has been a challenging one. Despite the fact that these systems seem to have many benefits to producers, evidence from the literature suggests that the adoption of decision support systems technology in the agriculture sector has been limited (Cox, 1996; Lynch et al., 2000; Hearn and Bange, 2002). Some of the reasons behind the low adoption of this technology include poor end user involvement, poor understanding of its potential benefits, complex operation of the system that requires considerable data input and computer skills, and outputs generated may not fit the producer's decision making style (Newman et al., 2000).
Traditionally, decision support systems have been developed primarily using standards of legacy applications, defined in this paper as those applications that have been inherited from languages, platforms, and techniques earlier than current technology of the Internet and the World Wide Web. Common consequences of legacy application development include software and database installation in users' personal computers, which sometimes may be a complex operation, as well as the need to update the software and/or input databases when these have been upgraded. Moreover, these legacy-type applications generally are not cross-platform compatible, so they are limited to users using the same computing platform under which these applications were developed. A Web-based application, on the other hand, would not require special installation in users' personal computers and would work across all computing platforms. Anyone with a computer connected to the Internet and a Web browser would be able to access and use it. Because of its centralized nature, application improvements and new developments would be instantly available to users without the need of software redistribution and reinstallations.
Newman et al. (2000) concluded that methods for executing simulation models via a Web interface had to be developed in an expedient manner to effectively deliver decision support systems to industry and policymakers. Advances in this field of technology transfer to farmers have been limited, but are increasing in number. Lassen et al. (2003) reported that during the years 1998 to 2002, the Danish Institute of Agricultural Sciences was involved in projects together with the Estonian, Latvian, Lithuanian, and Polish agricultural institutes to test, adapt, and implement for these countries a Danish Web-based decision support system for plant protection named Pl@nteInfo, developed by Jensen et al. (1998, 2000). More recent developments include, for example, the Rice Development Advisory http://beaumont.tamu.edu/RiceDevA/RiceDevA.aspx, cited 25 June 2005, verified 19 Jan. 2007), a Web-based program that predicts rice growth stages and provides advice on rice production; the High Plains Evapotranspiration Network http://amarillo2.tamu.edu/nppet/petnet1.htm, cited 25 June 2005, verified 19 Jan. 2007), a Web-based program focused on weather data collection and calculation of evapotranspiration; and WEATHERINFO, a Web-based system that focuses only on capturing weather data from other websites and deploying them to client sites (Steiner et al., 2005).
This paper discusses structural considerations regarding the development of Web-based decision support systems and describes the development and deployment of a decision support system that provides cotton producers in the Coastal Plains of Texas easy and rapid access to (a) current and historical weather data and (b) a suite of crop management tools entirely over the Internet using Web-based technology.
|
STRUCTURAL AND DEVELOPMENTAL CONSIDERATIONS |
|---|
|
TOPABSTRACTINTRODUCTIONSTRUCTURAL AND DEVELOPMENTAL...IMPLEMENTATION OF A WEB-BASED...CONCLUSIONSREFERENCES |
|---|
But the perceived ease of use is just one important feature in the design of decision support systems. The other important feature of decision support systems that will affect their adoption is their perceived usefulness. According to Keil et al. (1995), perceived ease of use is a measure of the reduction (or increase) in physical or mental effort to use the decision support system. Alternatively, perceived usefulness is a measure of how well the decision support system will enhance a user's decision-making capability. Keil et al. (1995) suggested that (i) software that rates low in ease-of-use and low in usefulness will be rejected, (ii) software that is high in ease-of-use and low in usefulness may be embraced by users initially but there is little chance of lasting acceptance, and (iii) software that is low in ease-of-use but high in usefulness will only be used by very competent computer users, while most users will avoid this type of software because the time and effort required to learn how to use it outweighs the potential benefit. Keil et al. (1995) concluded that the aim should be to develop software that rates high in ease-of-use and high in usefulness.
Structural Considerations
The most widely implemented approaches to deliver decision support systems are classified as data-driven and model-driven approaches (Bhargava and Power, 2001). Data-driven approaches help managers organize, retrieve, and synthesize large volumes of data using database queries and other techniques. Model-driven approaches, on the other hand, commonly provide analytical support using tools for decision analysis, statistics, simulation, logic, and stochastic modeling. The focus of the description presented below centers on the development of a Web-based decision support system that combines both these approaches. For the sake of simplicity, all Web-based systems discussed below refer to a combined data- and model-driven decision support system.
A carefully structured database system is perhaps the most important ingredient of a Web-based decision support system. The first step in planning the database system should place careful consideration to the different types of data being stored and retrieved within the system and secure the cohesive flow of all of them. In a typical Web-based decision support system the number of databases can easily reach into the hundreds to accommodate the data needs of applications, calculation tools, the system, and its users. Databases for storage of user data and stored results from applications, calculators, and tools will obviously need room to grow.
The second step in planning the database system must include careful consideration of what if scenarios. Web-based decision support system planners should visualize and plan for future changes in database configurations, including additions of new databases and increased number of users. The addition of databases and users make scalability a very important matter.
Third, increased numbers of both users and applications will eventually lead to increased load on hardware. Planning of a successful Web-based decision support system should include, therefore, careful consideration of hardware expansion. Hardware expansion will eventually become an issue in an open decision support system, as applications, users, and databases are destined to increase in size and number. Therefore, the system should permit easy expansion of hardware and seamless communication among added components that do not compromise the performance of databases and applications.
Software Development Platform
The choice of software development platform is critical when developing Web-based decision support systems because of the considerations regarding growth issues and because the Web-based software to be developed (server-side scripts or programs) must fit the Web-based model that is organized around HTML (HyperText Markup Language) documents and their contents.
The Web-based software development platform should be powerful, to accommodate a wide range of computational needs, and user-friendly, to make the development of programs easy. In the case of the development of open Web-based decision support systems, the software development platform should invite the participation of other software developers.
With regard to growth issues, the software development platform should permit the handling of numerous and large databases efficiently, enable easy communication among hardware components should the need for hardware expansion arise, allow the development of Web-based programs compatible with currently used client-side Web browsers, and preferably be independent from Web-server platforms.
|
IMPLEMENTATION OF A WEB-BASED DECISION SUPPORT SYSTEM: PUTTING DEVELOPMENT CONSIDERATIONS TO WORK |
|---|
|
TOPABSTRACTINTRODUCTIONSTRUCTURAL AND DEVELOPMENTAL...IMPLEMENTATION OF A WEB-BASED...CONCLUSIONSREFERENCES |
|---|
The following basic rules for the development of the Crop Weather Program were set at the start of the project: the decision support system should be easy to use, should be not intimidating but friendly and inviting to the end users, should be useful, should be computationally as well as scientifically robust, and should have a modular and expandable structure to facilitate future growth. The primary goal of the Crop Weather Program was to provide cotton growers easy access to (a) weather data collected by a network of weather stations and (b) a suite of powerful numerical simulation tools for field-specific calculations of a variety of state variables describing the crop and its environment, such as heat unit calculation, occurrence of phenological events, soil moisture content, canopy growth, and crop water use. The Crop Weather Program was launched in March 2000 (Fernandez et al., 2001) and delivered to crop managers through http://cwp.tamu.edu/ (verified 19 Jan. 2007). The Crop Weather Program is an expanding decision support system that integrates numerous smoothly interconnected Web-based applications and databases housed in two Linux-operated servers (Linux is an open-source version of the UNIX operating system).
HTML/OS (Aestiva, LLC, Torrance, CA) was the software platform chosen for developing the Crop Weather Program because (i) it is a purely Web-based system for software development, (ii) it integrates programming language and a database engine both powerful and easy to use, (iii) it has built-in networking capabilities that allow seamless and easy communication between servers (server jumping), (iv) it can be housed on the vast majority of today's popular server platforms, and (v) it is compatible with all contemporary Web browsers. HTML/OS reads standard HTML and allows adding programming code before, within, and/or after the standard HTML document code. Its programming language handles user-defined variables in the form of single values, tables, and text code (subroutines), and uses many predefined words (Otags) that perform specific operations common to word or text processing, spreadsheet or array processing, database management, and mathematical functions. HTML/OS also provides XML (eXtensible Markup Language) tags that allow to store, search, access, and utilize both XML and HTML data.
Since numerical crop simulation models require input of weather data, the first phase in developing the Crop Weather Program consisted in securing the automated acquisition and archiving of weather data and its easy access. The Crop Weather Program currently established a network of 22 automated weather stations to be used as source of weather data. Twenty of these weather stations are located in 12 counties of the Coastal Plains of Texas from Fort Bend to Kleberg counties. A completely automated smart system was built to retrieve, upload, and inspect weather data and update databases. A Microsoft Windows XP-operated computer located at TAMUAREC-Corpus Christi routinely connects to weather stations in the network on an hourly basis and downloads new data using the data logger support software LoggerNet version 2.1b (Campbell Scientific, Inc., Logan, UT). Weather data include air and soil temperature, relative humidity, solar radiation, wind speed and direction, and rainfall. The newly downloaded data are then transferred via the Internet to the Crop Weather Program's website, where a Web-based application program is launched to inspect the new data for integrity and update the weather databases. If the inspection reveals faulty data, such as values outside normal range and repeated or missing hourly records, the program alerts the website's administrator of the problem via an automated email message. Users accessing the Crop Weather Program's website can rapidly search hourly or daily weather data collected by any of the weather stations in the network. Weather data searches are displayed in downloadable tabular form.
Further developments of the Crop Weather Program focused on user registration, user login, field profile creation and management, and the development of numerical simulation tools for field-specific calculations of state variables describing the crop and its environment. Registration allows a user to define a unique username and password, with which is created a private password-protected partition of the Crop Weather Program website named MyCWP. A registered user can then log into MyCWP to access the suite of crop and weather calculation tools. To use a calculation tool, a registered user must create at least one field profile. A field profile is a set of unique information pertaining to a field management unit, such as the name of the field, year of production, soil type, slope of the field, type of water supply, planting date, planting density, row spacing, cultivar planted, and so forth. Field profiles are easily created using a very friendly user interface. Once created, a field profile is automatically stored in a database and its data used later as inputs for performing calculations, thus avoiding repetitive data entries every time a simulation tool is executed. Each user can create an unlimited number of field profiles. Numerical simulation tools available in MyCWP include Reference PET, Crop Development, Pre-Planting Soil Temperature, Post-Planting Soil Temperature, Soil Moisture, Crop Water Use, Irrigation Monitor, and Defoliation. These simulation tools are very easy to use and produce information useful to crop managers; they have been developed based on the needs of cotton producers. Outputs generated by these programs are displayed in easy-to-read, uncomplicated summary tables and charts. Charts are created online in real-time with the third-party software FusionCharts (InfoSoft, Kolkata, India). Tabular display of more detailed day-to-day calculation results is optional. Users are also given the option of downloading these day-to-day calculation results. A brief description follows of the currently available tools.
The Reference PET tool is not specific for cotton and calculates reference potential evapotranspiration and cumulative rainfall (current and historical average) during a specified period of time. The objective of this tool is to provide a quick comparison between atmospheric evaporative demand (potential water loss) and rainfall (natural water supply). Reference potential evapotranspiration is calculated at hourly steps using the Penman–Monteith equation and applying the reference (standard) method described by Pereira et al. (1996). Reference evapotranspiration has been defined by Allen et al. (1994a, 1994b) as the rate of evapotranspiration from an hypothetical reference crop with a height of 0.12 m, a fixed canopy resistance to water vapor flow of 70 s m–1, and an albedo of 0.23, which closely resembles the evapotranspiration of an extensive surface of green grass with uniform height, fully shading the ground, and with nonlimiting water supply.
The Pre-Planting Soil Temperature tool allows monitoring the progression of the minimum soil temperature during the previous 10 d of a selected date, thus providing useful information for determining proper time to initiate planting of cotton. Two soil temperature indexes are provided. The first index shows the progression of the 10-d running average of the minimum soil temperature at 0.203 m (8 inches) and provides information usable by followers of the rule-of-thumb proposed by Holekamp et al. (1960). The second index shows the progression of the daily minimum soil temperature at 0.076 m (3 inches) and provides soil temperature information usable by followers of the general recommendations for planting cotton in Texas (Sansone et al., 2002). Outputs produced by the first index are more stable than those produced by the second index since they are the result of 10-d running averages instead of daily values, and are based on soil temperatures measured at a depth of 0.203 m (8 inches) instead of 0.076 m (3 inches). Amplitudes of diurnal and seasonal soil temperature waves decrease with depth (Rosemberg, 1974).
The Post-Planting Soil Temperature tool allows monitoring the progression of daily minimum soil temperature at a depth of 0.025 m (1 inch) after sowing, thus providing useful information on the thermic environment surrounding the seed during germination and emergence for management decisions such as those involving replanting. The minimum temperature for germination of cotton is about 12°C (53.6°F) (Arndt, 1937). The occurrence of temperatures below this level during germination and emergence can cause injury to cotton. The immediate effects of low soil temperatures during germination are radicle tip abortion and root cortex disintegration (Christiansen, 1963). Emerged seedlings are also sensitive to soil temperatures below 12°C (53.6°F), which can cause inactivation of root water uptake, leading to seedling desiccation (Christiansen, 1979). The effects of low soil temperatures during germination can also cause long-term growth reduction and delay in flowering (Christiansen and Thomas, 1969). Outputs generated by the Post-Planting Soil Temperature tool are shown in a chart that displays the progression of daily minimum soil temperature at a depth of 0.025 m (1 inch) during the 15 d that follows planting, as recorded by the weather station specified by the user in the field profile.
The Crop Development tool estimates the progression of phenological stages in cotton under current weather conditions and historical average conditions. Prediction of crop development phases provides useful information for timing activities such as side-dress fertilization, irrigation, and use of plant growth regulators and insecticides. The date of a particular phenological stage is predicted as a function of cultivar maturity, heat unit requirement, and postplanting cumulative DD60 (degree-days > 60°F or 15.6°C), applying a set of empirical equations specially developed for this application that relate DD60 accumulation and occurrence of phenological stages (Fernandez, unpublished data, 2002). For developing this set of empirical equations, an initial set of empirical equations was obtained first using data collected for ‘Deltapine 50’ at Corpus Christi, TX, from 1989 to 1996 (J.A. Landivar, 1998, personal communication). A computational algorithm was then created to apply this initial set of equations to cultivars of different maturity. The accuracy of this online tool for estimating phenological dates was confirmed by the highly significant regression of observed dates on estimated dates (y = 0.9736x + 5.4581, r2 = 0.9718, significant at P = 0.01). This regression equation was obtained using data collected from an irrigated test that included 22 cultivars in 2001 and a rain-fed test that included 24 cultivars in 2002.
The Defoliation tool uses observed date of the "five nodes above white flower" (5NAWF) stage and/or NAWF number to calculate the DD60 accumulated after the 5NAWF stage. The 5NAWF stage has been defined as the stage when cotton plants produce the last effective flower population, also defined as cutout (Bourland et al., 1992). The Defoliation tool predicts dates of defoliation of upland cotton based on accumulations of 850, 950, or 1050 heat units after 5NAWF (Young et al., 1980; Wells, 1991; Benson et al., 2000; Witten et al., 2001).
The Soil Moisture tool applies only to fallow fields and provides users with valuable information regarding the progression of soil moisture storage and distribution of moisture throughout the soil profile. Estimates of the moisture content throughout the soil profile before planting can be useful for management decisions related to seeding rates, row distance, cultivar selection, and fertilization rates. New algorithms were developed for Soil Moisture to simulate soil evaporation, rainfall/irrigation water infiltration, and soil water balance at hourly time intervals (Fernandez, unpublished data, 2002). Soil evaporation is calculated using the Penman–Monteith equation adapted to a bare soil surface and the newly developed algorithms to simulate the effect of soil drying on the rate of soil evaporation. Rainfall water infiltration is calculated using an algorithm developed to take into account soil infiltration properties, rate of precipitation, and moisture of the upper layers of the soil. The soil water balance in each 0.025-m (1-inch) soil layer is calculated using a simple book keeping method that takes into account current balance, gains (rainfall and irrigation), losses (soil evaporation), and maximum water holding capacity. This simple book keeping method is applied in cascade to all layers of the soil profile from top to bottom. Soil physical characteristics used for simulating the soil water balance are retrieved from the Crop Weather Program's soils database using the soil type specified in the user's field profile. This soils database was built using data obtained from soil surveys by county published by the USDA/Soil Conservation Service (currently Natural Resources Conservation Service, NRCS) in cooperation with the Texas Agricultural Experiment Station and currently also with the Texas State Soil and Water Conservation Board, as well as from the Soil Survey Geographic database for various Texas counties published by USDA, NRCS, and the National Cooperative Soil Survey. Information retrieved from the soils database includes soil name, layer identification, soil parameters describing layer depth, available water capacity, bulk density, and permeability.
The Crop Water Use tool simulates the progression of crop growth and canopy development, crop water use, soil moisture storage throughout the soil profile, and cumulative soil water deficit at root depth. Once this application tool is launched, calculations are performed at hourly steps. The Crop Water Use program is organized around three main components. The first main component calculates the development of the canopy in terms of height, leaf area index, and groundcover using a series of empirical equations relating main-stem plastochron and internode elongation to air temperature (Reddy et al., 1997) and plant height to leaf area index (Marani and Ephrath, 1985). New algorithms were developed to relate expansive growth to soil water content available to the plant (Fernandez, unpublished data, 2002). Soil volume occupied by roots is calculated as a function of canopy growth. The second main component calculates water fluxes from the soil and the canopy separately. Algorithms used to calculate soil evaporation are similar to the ones used in the Soil Moisture tool. Two soil evaporation fluxes are calculated, one for sunlit soil and the other for soil shaded by the canopy. Calculation of canopy transpiration is based on the Penman–Monteith method, where canopy height is a variable and the canopy resistance to water flow is a function of plant-available soil water content (McCree and Fernandez, 1989). Crop water use is calculated as the sum of sunlit soil evaporation, shaded soil evaporation, and canopy transpiration prorated on ground cover. The third main component calculates the soil water balance at 0.025-m (1-inch) increments using a procedure similar to the one described above for the Soil Moisture tool. Two soil water balances are calculated, one for nonrooted soil and another for the rooted soil. The accuracy of the computational algorithms used in Crop Water Use tool has been preliminary confirmed in a recent study (Medeiros et al., 2006) that showed good agreement between calculated values of plant height, leaf area index, and total soil moisture content and their corresponding observed experimental data collected from irrigated cotton crops at a location near Halfway, TX.
The Irrigation Monitor tool (a modification of the Crop Water Use tool) allows irrigators to estimate how the crop uses the last applied irrigation water and any additional rainfall amount, thus providing useful information for irrigation scheduling and/or the adjustment of irrigation systems. This tool applies only to cotton and is usable only when irrigation records have been entered in advance using the My Irrigation online tool available at My Field Observations partition of MyCWP. Irrigation records are archived in a database and each record consists of the date of an irrigation event, the amount of water applied, a discrete qualitative description of the stress condition of the crop (selected from a dropdown menu of options), and an average value of plant height at the time of irrigation. This information is then used to generate the needed inputs to run the Irrigation Monitor program. The main difference between Irrigation Monitor and Crop Water Use is the time frame for performing calculations. While Crop Water Use starts from planting date, Irrigation Monitor starts from the date of last irrigation. The most important output generated by Irrigation Monitor is the amount of water (irrigation and rainfall) left available to the crop since the last irrigation event expressed as percentage. Irrigation managers can use the Irrigation Monitor to schedule the application of the next irrigation.
Currently, the Crop Weather Program includes >40 application programs that access >500 subprograms in the form of dynamic HTML, text, and XML files, and >190 databases. About three-fourths of the application programs are components of the suite of simulation and calculation tools available to registered users. The rest of the application programs are utilities exclusively used in development or administration of the Crop Weather Program. Most of the databases are weather-related (five per each of the weather stations in the network), the majority of which are updated hourly. Other databases contain information relative to the clientele of users, user's field profiles, the weather station network, soils descriptions obtained from the Soil Survey Geographic database (published by USDA, Natural Resources Conservation Service, and National Cooperative Soil Survey), commercial cultivars available and their characteristics as described by seed companies, user's field observations on crop growth and development, applied irrigation, and field-measured rainfall, outputs from previous user's simulation runs, temperature and precipitation normals, and so forth. Operation of the Crop Weather Program involves intense server jumping which goes unnoticed by the user, as main programs residing in one server may access databases and other support programs residing in the other server. Additional hardware expansions are foreseen, occurring as the number of registered users increase, the network of weather stations is expanded, and new numerical simulation tools and utility programs are developed.
The usefulness and easiness-of-use of the Crop Weather Program has been reported in popular media (Lawrence, 2001; Buehring, 2001; Ramirez, 2001; Smith, 2005), but these reports were based only on information gathered from a limited number of users. Records in the Crop Weather Program's user database shows that there are currently >300 registered users who created >750 field profiles. Login information obtained with WebTrends Log Analyzer (WebTrends Corporation, Portland, OR) from March 2003 to February 2005 shows that there were 1 572 318 successful hits to the entire website, which included 48 958 visitor sessions. About 37% of the visitor sessions, however, can be attributed to weather data transfers, database updates, and program development. Although these statistics may seem impressive at first glance, they do not provide enough information to properly evaluate how well this Web-based decision support system is being used. A tracking system embedded in each of the main individual components is needed to collect the information required for a proper evaluation of the usage of the Crop Weather Program. Development of such a tracking system and a parallel system to collect feedback information from users are currently in progress.
The vast majority of the limited resources available for this project were devoted to software development, and only a small portion of them were directed to promotion and training. Annual hands-on online workshops and occasional one-on-one training sessions with users facilitated the gathering of valuable feedback information that not only inspired new developments but also revealed a short user learning curve and their willingness to adopt this technology transfer system—A confirmation that the Crop Weather Program is an easy-to-use and useful Web-based decision support system.
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONSTRUCTURAL AND DEVELOPMENTAL...IMPLEMENTATION OF A WEB-BASED...CONCLUSIONSREFERENCES |
|---|
Main current limitations of the Crop Weather Program for South Texas include (i) usability limited mainly to the Coastal Plains region of Texas, as this is determined by the extent of the network of weather stations established by the program; (ii) sole applicability to cotton, since Cotton Incorporated was the main funding source; (iii) unavailability of tools for predicting yield and economic returns; and (iv) unavailability of a well-developed method for gathering and analyzing usage and feedback information from users. A simple and automated method for expanding the network of weather stations to other cropping areas and/or for allowing users to input weather data sets needs to be developed to broaden its geographical applicability. Also, new modular applications need to be developed and added to the suite of available tools to assist the decision-making process that relates not only to cotton production but to other crops as well, preferably with involvement of outside developers. A better and efficient method for gathering usage and feedback information needs to be developed to identify areas in the current system in need of adjustment and guide future developments. The concept embedded in the Crop Weather Program for South Texas is consistent with and open to a multidisciplinary development effort, and its modular and scalable architectural design facilitates the addition of a wide variety of new tools.
Also, after >6 yr, HTML/OS continues to deliver a reliable and rapid software development platform for this project. Since the Crop Weather Program development plan anticipated expansion of the network of weather stations, its user base, and the number of applications served, the built-in networking capabilities of HTML/OS played a crucial role by allowing uncomplicated modular expansion of the system hardware.
We anticipate that this type of approach for technology transfer to farmers using Web-based decision support systems will continue to expand worldwide and play a crucial role in farming since computing is prevalent in everyday life and numerical simulation provides a cost-effective and attractive means to quantify biological and physical processes that interactively determine the outcome of crop production.
|
ACKNOWLEDGMENTS |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONSTRUCTURAL AND DEVELOPMENTAL...IMPLEMENTATION OF A WEB-BASED...CONCLUSIONSREFERENCES |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| The SCI Journals | Crop Science | Vadose Zone Journal | |||
| Journal of Natural Resources and Life Sciences Education |
Soil Science Society of America Journal | ||||
| Journal of Plant Registrations | Journal of Environmental Quality |
The Plant Genome | |||
