|
|
A Case Study in Phrao District, Northern Thailand
Conclusions
Below the main conclusions are presented per main objective of the case study (see chapter 1).
Objective 1 (planning process):
| 1. | The process of identification of required data, data collection, assembly, analysis and scenario development for this case study took approximately four man months. The required information was widely scattered over different agencies, so collection of the data was a time consuming process, partly due to the unfamiliarity of the author with the agencies. Now data sources and agencies have been identified, repetition of the data collection and analysis to obtain updated data should require considerably less time (± one man month). To conduct a similar study for another district would approximately take two months since a lot of common data sources have already been identified. Considering the present institutional set-up (i.e. centralised planning, weak local agencies) it cannot be expected that similar studies can be conducted for the whole country.
|
| 2. | Although data are incomplete and sometimes lacking, assessment of the current energy situation, scenario development and forecasting for decentralised energy planning are still possible and feasible with respect to data availability. A start can be made with available data to identify data gaps, collect additional data, repeat analysis, identify data gaps, etc. Lacking data can be supplemented by data from other areas and by reasonable assumptions based on available data and some background information on the area and related issues; |
Objective 2 (data):
| 3. | Data had to be obtained from a wide variety of agencies at different administrative levels, mostly government agencies. Consequently the data were often difficult to integrate because they differed in definitions, scale, period and stratification. It was found that at present there is virtually no co-operation between agencies with respect to data and that agencies are often not aware of data collected and published by other agencies that may be useful for their own purposes. More co-ordination and co-operation is required to develop standards for data collection and presentation and to allow for the processing of raw survey data for specific purposes.
|
| 4. | Data came mostly from government agencies, such as NSO, DEDP, DLD, RFD and the district office. Few data were identified from universities and private organisations. However, more data may be available but time was lacking to check this. At central level data were provided by several divisions or sections of five government agencies (CDD, DLD, DEDP, NSO, RFD). Four provincial offices provided data (RFD, PSO, PIO, OAC), while at district level the district office could provide data on several issues.
|
| 5. | The National Statistical Office conducts surveys on a wide variety of topics on a regular basis and possesses a large amount of data. However, it was found that these data are not always useful because in several cases they are not consistent with other NSO surveys or with data from other sources regarding definitions and others. For example, boundaries of household income groups of subsequent energy consumption surveys do not correspond with each other, which complicates the historical analysis of consumption patterns, and these boundaries also do not correspond with data on the number of households per income group per district as published by NSO. Stronger co-ordination between NSO surveys on the one hand and between NSO and other agencies dealing with specific topics on the other hand may enhance the usefulness of data for planning. NSO could play a central role in this, since its primary task is data collection and publication. Naturally, this would not only benefit the energy sector. This may apply to statistical offices in other countries as well.
|
| 6. | The lack of data is often mentioned as a constraint to the implementation of energy planning in rural areas at national and sub-national level. A comprehensive data base is often considered as a prerequisite for energy planning. The case study has shown that a lot of data do exist and that the energy assessment at district level is feasible in the case of Thailand. Data base development occurred in several stages parallel to energy demand and supply analysis through the identification of missing or insufficient data. This showed that data base development is not a prerequisite to energy planning, but that it is part of the continuous process of energy analysis and assessment. |
| 7. | Data uncertainty can have a strong impact on the modelling results. Although for most data the uncertainty cannot be assessed analytically because a series of observations cannot be made or are unavailable, the uncertainty can be assessed by the estimation of the most probable range for each variable, and by the evaluation of error propagation. The impact of assumptions on driving parameters, such as growth rates, macro-drivers and elasticities can be evaluated by the use of sensitivity analysis. Before a new stage of data collection, analysis and scenario development, that should focus on those factors that cause the highest uncertainty, is conducted, the benefits of additional data should be evaluated compared to the cost of obtaining these data. If the costs are higher than the benefits, other methods should be identified to evaluate the need for and the type of interventions.
|
| 8. | Time series data are often inconsistent with respect to definitions, coverage, scale, period and other factors which complicate their application to planning; trends are difficult to identify because changes may appear more due to different methods applied than any actual change that has taken place. Consistency of time series data may be more important for planning purposes than higher accuracy or a larger coverage. |
| 9. | In order to be useful for planning, data should distinguish different groups, areas or others to allow for analysis and forecasting. Especially for household energy this is important because households are small units and their behaviour shows a high variation. It is relevant to study the relationship between, for example, energy consumption and income (preferably gender-disaggregated), type of area (rural, sanitary, urban), household size, and occupancy. Forecasts can be based on (assumed) shifts in the change of population per group or area or the change of behaviour towards the behaviour of another group or area. Also a distinction per type of environment (e.g. agro-ecological zone) would provide data that are useful for planning. |
| 10. | The case study confirms that household energy consumption is site-specific and actually requires local surveys. The DEDP and NSO surveys for the Northern Region show large differences with the household survey in Phrao, especially in respect of fuelwood and charcoal consumption. In Phrao woodfuel use is higher and charcoal use lower than the average consumption for the Northern Region, probably because of the wood abundance in the district. Naturally, local surveys are more cost- and labour-intensive and time-consuming which prevents the coverage of all districts in a similar way as for Phrao. However, if data from large-scale surveys distinguished different user groups, areas and others, there would be less need for local surveys (see point 9).
|
| 11. | Although the ITC reports provided useful information, most of it was not essential for the case study. Only the household energy survey (Arriola, 1993) provided valuable information since household energy consumption is site-specific (see also point 10). |
| 12. | On the energy demand side enough data are available, but on the biomass supply side some data are lacking, especially on biomass resources (stock and yield). For forest areas, data of other areas and countries can be used since these are relatively homogeneous and comparable. Although for the non-forest wood resources in Phrao the same method was used, this is not realistic. More reliable methods are needed to assess these resources.
|
Back to the top
Objective 3 (minimal approach):
| 13. | To limit the time and skill requirements a minimal approach to data collection and analysis can be followed. Of course this is related to the level of decentralisation at which planning is actually conducted, but considering the present institutional set-up in Thailand this is most feasible at central level. Feasible options to minimise data collection and analysis are the use of central data only, restriction to major sectors, the use of simple forecasting methods, and accessibility assessment without spatial evaluation.
|
Objective 4 (planning tools):
| 14. | The Long-range Energy Alternative Planning model (LEAP) uses a flexible data structure and it includes a biomass module for the evaluation of the resources of wood, agriculture residues and dung, all of which make it applicable for planning at area-based level. It is appropriate for the method of iterative data base development as applied in the case study. It provides a comprehensive framework for the covering of the whole energy flow from biomass resources through conversion to end-use consumption. Forecasts for single factors can be made by using growth rates, macro-drivers and elasticities, or by explicitly specifying a value for each data year. In the last way LEAP can easily incorporate the results of other models that focus on a specific part of the energy flow or that apply another method, e.g. econometric demand models or simulation of power plants.
|
| 15. | Before data could be entered into the model, the available data had to be analysed and converted into an appropriate format for LEAP. For this additional tools were required. Especially spreadsheets (such as Lotus 1-2-3, Quattro Pro, and Excel) can be useful and they were used extensively during the case study, for example to assemble data in tables, to convert energy consumption values using standard heating values for fuels, and to make population forecasts. They were also used for the formatting and presentation of results such as energy balances.
|
| 16. | For the evaluation of the resources of wood and other biomass in the study area, LEAP's biomass module combines data on land use with data on tree stock and yield and crop productivity per land use type. Data on the area, stock and yield, and productivity have to be specified per land use type. Land use forecasting can be incorporated by specifying future conversions of parts of one land use type to another, e.g. agricultural land to settlement area. The accessibility of wood resources is incorporated by an access fraction for each land use type in which several constraining factors can be combined. This fraction determines the area, and thus the amount of woody biomass, that is accessible for each land use type. The model can incorporate all of the above factors but it provides no tools or methods to obtain the required input data, so data analysis has to occur externally. |
| 17. | The Environmental Data Base (EDB) of LEAP contains data on emissions of materials to the atmosphere, water and soil for typical end-use and transformation devices, such as cookstoves, boilers and charcoal kilns. These data have been obtained from various literature sources reporting research on the emissions per unit of consumed fuel for the devices. The EDB can be used to evaluate the emissions for alternative scenarios. However, a major shortcoming is that it lacks an adequate description of the end-use and transformation devices that are included in the EDB. This complicates the linking of the devices in the scenarios with the devices of the EDB and limits the usefulness of the EDB since the results depend largely on the selection of the devices. An adequate description of the devices will enhance the usefulness of the EDB and the environmental impact assessment. |
| 18. | A Geographic Information System (GIS) can be a useful tool to analyse the available spatial data and to produce the data required for LEAP. Spatial land use data can be converted to attribute data, and land use changes can be evaluated by the development of (conceptual) models that describe spatial and non-spatial relationships of land use types. Furthermore, accessibility can be evaluated spatially, as was done in the case study. When land use information is not available or up-to-date, satellite images can be useful to obtain the required data. Because experience and equipment will mostly be lacking at energy departments, cooperation with other departments that do have the skills and facilities is required (e.g. forestry).
|
Back to the top
Table of Contents
|
