The RAISON GIS (regional analysis by intelligent systems on microcomputers geographic information system) was designed to manage the microbiological quality of rural drinking water sources, by providing an efficient means for the storage, analysis, and presentation of a large volume of water-quality monitoring data.
The International Development Research Centre (IDRC) sponsored the development of this information system at the National Water Research Institute in Burlington, Ontario, Canada (Tong et al. 1989), for the purpose of data collection and processing in and by both developed and emerging nations. First used to assess the risks posed by acid rain to water sources in eastern Canada (Lam 1986), the RAISON system is an integrated package incorporating a database management system, a mapping package, and a spreadsheet. It utilizes a set of GIS programs under the RAISON Programming Language environment for a multilevel, menu-driven presentation of water quality data and related information in map forms.
Based on the RAISON system, a prototype system called µRAISON was developed at the University of Malaya as part of a rural water supply, sanitation, and drinking water surveillance programme. µRAISON uses additional common software packages for certain parts of its implementation, such as dBASE III for complex database management, Autocad or Crosstalk for map digitization, and SPSS PC+ for statistical analysis. A separate database management package was also employed in view of the versatility of dBASE III.
The second task involves the creation of various database files, in which geographic locations are identified by state, district, village, or station names. Statistical information on water supply, sanitation facilities, and other relevant socioeconomic information for the respective geographic levels is entered directly into the RAISON database. However, to facilitate a more direct access for statistical analysis, water quality monitoring data and other complex sanitary survey features have been set up in dBASE III.
The third task is to develop the programs in the RAISON Programming Language environment so as to allow interactive menu-driven data analysis and presentation. Complex steps involving database and spreadsheet operations linking with GIS presentations are transparent to the users.
µRAISON System Configuration
The setting up of tasks of the µRAISON Water Quality Data Management System involves
map digitization and the creation of database files. Updating could be carried out externally. However,
statistical analysis is carried out separately from the µRAISON package.
Map digitization uses packages such as Crosstalk or Autocad for the generation of vector graphic files. The files are then converted into map files for incorporation into RAISON using the "calibration/conversion programs" developed for µRAISON.
Water quality and sanitary survey databases are created in the dBASE III+ environment and then exported to the main RAISON database for source classification and map presentation. Within the µRAISON system, procedures have also been developed for exporting these database files to SPSS PC+ for statistical analysis.
Two additional options are also available for µRAISON after system startup, namely :
µRAISON Users
The µRAISON system was targeted for two levels of users. Using the menu-driven GIS options,
an operator with minimum computer knowledge can update data and execute specific data
analysis and GIS presentations. Advanced users can follow the user manual to add or change
functions in the data analysis and map presentation to meet their specific needs for management
and planning.
Sanitary Survey and Water Quality Data
More than 1500 pieces of field testing data on the microbial water quality of some 800 water
sources have been compiled. These have been stored in dBASE III+ files and imported into the
µRAISON database to be utilized for testing, programing for analysis, and display within the
µRAISON system environment. A total of 15 village/region map files have been incorporated.
These include four villages each in the states of Selangor and Negeri Sembilan, one village in Pahang, three regions in Kedah, and three villages in Kelantan.
Sanitary survey records for the sampling stations were also obtained at the time of sampling, and a database file designed for their storage. The database file consists of four sections: type of water supply; sanitary protection of the water source; sources of pollution; and land usage.
The data gathered on water supplies and sanitary surveys provided such information as: type of well (e.g., dug, tube, pipe, public water supply, open watercourse); state of repair (as evidenced by cracking or crumbling casing or presence of rubbish); surrounding ground type (clay, sand, etc.); population density; presence of small children; and agricultural activity (particularly the presence of animals). In addition, data on type, availability, and status of adjacent latrines were included. These data are relevant as many examples of well contamination can be related to the presence of latrines which are inefficient due to their flushing system, the large number of individuals using them, their proximity to wells or gradient from wells (upslope or downslope), and their general state of repair/cleanliness.
Subjective Classification System Based on Commonly Adopted Standards
Fecal coliforms have been widely accepted as the indicator organisms of choice for the detection
of human fecal contamination of water sources. Wang et al. (1989) have also demonstrated the
usefulness of coliphage enumeration for this purpose.
Data obtained on coliphage and m-FC (membrane fecal coliforms) counts have been used directly for the classification of the rural water sources. Ranges of coliphage or m-FC counts of < 5, 5-50, 50-250, 250-1,000, and >1,000 per 100 mL sample have been classified as Class I, II, III, IV, and V, respectively. A systematic colour code has been adopted in the µRAISON GIS for the display of water sources quality in map format. Class I quality is consistent with the WHO guidelines for drinking water (1984). Classes II and III represent moderate quality in which water requires pretreatment such as boiling or chlorination before drinking. Classes IV and V are poor quality water of high risks which called for remedial actions on the water sources.
The usefulness of this classification scheme can be judged on the basis of the data sets tested. The classification results obtained for both m-FC and coliphage data for the 1500 water sources in 15 villages/regions show excellent consistency.
Objective Classification Using a Single Parameter
An objective classification system for demonstrating the quality of rural water sources can be
obtained using a single parameter that follows a negative binomial distribution based on the
estimation of the index of dispersion for the observed data (Shaarawi et al. 1981). The coliphage
or m-FC has been chosen as the single parameter to be classified separately .
The analytical results showed that if data on very high coliphage counts (>240 counts/20 mL or 1200 counts/100 mL) were not included in the fitting, the coliphage data follow a negative binomial distribution.
The group classification obtained based on coliphage counts was as follows:
| Counts/20 mL | Counts/100 mL | |
| Class I | < 1 | < 5 |
| Class II | 1-10 | 5-50 |
| Class III | 10-31 | 50-155 |
| Class IV | 31-62 | 151-310 |
| Class V | > 62 | > 310 |
More classes were derived if data of higher coliphage counts were included, although the ranges of lower groups were not affected. It is reasonable to group all data values > 62 counts/20 mL (> 310 counts/100 mL) as the highest group, due to the small number of samples observed over the range. It is noted that the classification results are comparable to those obtained using the subjective method.
Ranking Method Using Several Parameters
A ranking procedure taking microbial counts and sanitary protection conditions of water sources
into account has been incorporated into the GIS options. The sanitary ranking for each water
source has taken into account the source protection, proximity of polluting sources, and soil type.
Preliminary statistical correlational tests were performed between the microbial counts and the sanitary protection conditions, using SPSS PC+ to define the ranking for water sources of high quality (class I) to high risk (class V). Statistical analysis was carried out for (a) type of well, (b) depth of well, and (c) well protection, versus the different ranges of m-FC or coliphage counts. All six cases showed significant correlation at the 0.1% level between the microbial counts and the sanitary conditions.
Multivariate Classification Based on Discriminant Analysis
Tong et al. (1990) have presented a multivariate classification scheme based on the discriminant
analysis for rural potable water sources. The multivariate classification approach is based on
discriminant analysis which correlates sanitary protection, polluting sources, land use information,
as well as field data on microbiological quality. Tong et al. have shown that a linear combination
model using the sanitary condition data correctly classified 78% and 73% respectively of the
m-FC and coliphage data into two distinct groups each (< 5 and > 5 counts/100 mL). The
corresponding percentage of correct classification was 70% when the m-FC counts were divided
into three groups (< 15, 15-500, > 500 counts/100 mL). Similarly, the corresponding percentage
of correct classification was 66% when the coliphage counts were considered in three groups (< 5,
5-250 and > 250 counts/100 mL).
Back to Contents -
To next section
This file was created 23 February 1996