Status of Digital Soil Mapping in the Bureau of Soils and Water Management
Paper presented during the GlobalSoilMap.net East Asia Node Conference – Advancing the Science and Technology of Soil Information in Asia: Launch of the Global Soil Partnership’s Asia Soil Science Network and its East Asia Node, 8‐11 February 2012, Nanjing, China
It is still classical soil mapping for routine soil survey and classification activities in the Philippines. Nevertheless, the Bureau of Soils and Water Management (BSWM) has the necessary equipment for Geographic Information System (GIS) and Remote Sensing (RS) as well trained manpower complement to conduct digital soil mapping presently done at research and project level basis. Our current digital soil mapping research is focused on the mapping of land attributes of the watersheds of Cabulig, Misamis Oriental and of Inabanga, Bohol through a research grant from the Australian Centre for International Agricultural Research (ACIAR). With three months of field soil survey work to commence this February 2012, the research will develop methods of rapid soil measurement that allow quick and inexpensive (though less precise) measurements/ estimates of important soil attributes. This recognizes that, in the face of soil variability, the appropriate strategy is to make lots of inexpensive measurements rather than a few, precise and more expensive ones. Rapid soil measurement methods include the use of proximal sensors, such as visible/near infra-red and/or mid infra-red spectrometers, and pedotransfer functions to estimate soil attributes. These methods still require the use of conventional laboratory and field measurements on a proportion of sample locations, to provide local calibration. We also have FAO-funded land degradation assessment to harmonize our digital data holdings with global land degradation data. Furthermore, we are in the process of acquiring satellite imageries to cover the whole country and once processed, re-issue all our maps based on orthoimages. While digital soil mapping may still be a long way to go for BSWM, nevertheless, this seems the only way to go as we are faced with hiring moratorium which when once lifted, we are faced only with computer literates but classical soil survey illiterates in the job market. We thank the organizers of this East Asia Node Conference for enabling us to participate in the partnership.
The demand for soil information in the Philippines today is much greater than it was decades ago. The need for soil data as input to policy formulation, rural development planning, and researches is no longer confined to the traditional fields of agriculture, architecture and construction. The environmental science field for instance, as a matter of national policy, requires Environmental Impact Assessment and proposed set of environmental protection management practices for major projects prior to issuance of an Environmental Compliance Certificate in order to promote mandatory self-monitoring and compliance to environmental standards. Soils is the third of three important environmental pollution that includes also air and water that is given special attention considering its vital role in waste management and nutrient recycling. The field of meteorological sciences is now focused with climate change modelling and soils, especially its role in regulating hydrological cycle is brought to the fore.
The Bureau of Soils and Water Management, a staff bureau of the Department of Agriculture, is the sole nationally mandated government agency on soils. Its activities encompass all the sub-disciplines of soil science; with its water mandate limited to water as a vital agricultural resource. The Bureau’s soil survey and classification activities dates back as early as 1903 when an American soil scientist, Mr. Clarence Dorsey, conducted the soil survey of Batangas province. The actual inventory of soils in the country started in 1934 when the Soil Survey Committee was organized by then Secretary of Agriculture and Commerce. This inventory was carried out in each and every province but because of limited laboratory facilities and available technology during those periods, the reconnaissance type of soil survey based on soil profile observation was used. Most of the laboratory analyses were mainly for soil fertility assessment. The profile observations recorded the key characteristics of the soil series, the key pedological unit used for mapping the soils of the province.
The soil survey work was briefly interrupted by World War II and it was not until after the war in 1945 that the Soil Survey Division was organized. It was in June 5, 1951 that the Congress of the Philippines enacted Republic Act No. 622 organizing the Bureau of Soil Conservation. Reconnaissance provincial soil survey activities continued. The reports were published in scales ranging from 1:75,000 (Cavite) to 1:250,000 in general while Palawan, a big province, was mapped at exploratory scale of 1:500,000. This variation in map publication scale was done to conform with the actual size and shape of the individual provinces, especially the island provinces. In 2011, we just celebrated our 60th Anniversary as the national soil resource agency.
In 1964, the Bureau was renamed Bureau of Soils with regional as well as provincial soil districts established to bring the services closer to the farming communities. The reconnaissance soil surveys of the 75 provinces that then constituted the Philippines were completed sometime in mid-1960’s. A total of 348 soil series were mapped and identified. There is lack of pedological descriptions and laboratory analyses of the representative pedons which made these early efforts rather difficult to link with the USDA Soil Taxonomy system of classification that the Bureau later on adopted. Nevertheless, being the only document for Philippine soils, these studies are the most important source of information in the province and the only credible soil maps for use in the various agricultural development planning and studies.
The succeeding decade of the 70’s was characterized by detailed soil surveys at 1:10,000 scale mostly carried out for irrigation development projects with assistance from the Food and Agriculture Organization (FAO) of the United Nations Organization. This also marked the first major attempt to use equivalent USDA Soil Taxonomy classification for each of the soil series. The surveys, however, were mostly confined to alluvial flooded lands covering four major irrigation projects mostly in Central Luzon, with a total area of about 152,000 hectares and 59 new soil series were identified.
The succeeding decades of the 1980’s and 1990’s were semi-detailed soil surveys with map scale of 1:50,000. About 20 provinces with 5,496,690 hectares were surveyed. The Bureau’s history mirrored the tumult that rocked the Philippines in mid-1980 as Corazon C. Aquino was swept into presidency by the People Power Revolution. In January 30, 1987, the Bureau of Soils was reorganized into the Bureau of Soils and Water Management retaining its staff function of soil resources survey, agricultural land resources evaluation, conservation, and research but its mandate was broadened to include the development and management of water resources through the construction of small water impounding systems, the promotion of shallow tube wells, and other water resources technologies to alleviate the impact of prolonged drought on standing crops. Included in its broadened mandate is artificial rain making or the conduct of cloud seeding sorties over areas suffering from seasonal aridity. The 1990’s was dominated by two five-year technical cooperation with the Japan International Cooperation Agency (JICA) which improved the capability of BSWM to provide soil analytical data and ushered a computer-assisted map digitization and spatial analyses through the establishment of Geographic Information System (GIS) and Remote Sensing laboratories. The JICA technical cooperation extended for another five years until 2005.
The first decade of the new millennium continues with municipal-level updating of soil resources as a major component in the delineation of our Strategic Agriculture and Fisheries Development Zones (SAFDZ) as required by Republic Act 8435 known as the Agriculture and Fisheries Modernization Act. The SAFDZ is to be integrated to Comprehensive Land Use Plan (CLUP) of the local government unit. It remains the current focus of its soil survey and classification mapping to these days.
Soil Survey Activities in BSWM Today
BSWM is still in traditional soil survey.
That is despite the presence of Geographic Information System (GIS) and Remote Sensing (RS) facilities and competent staff that have undertaken graduate studies here and abroad, with one completing his PhD in Australia just this 2011. Since the introduction of state-of-the-art GIS and RS facilities in BSWM by JICA in 1991, we have upgraded our equipment including softwares four times – in 1997 through technical cooperation project with JICA as we migrated from mainframe to desktop computing and established our local area network, in 2003 also through technical cooperation project with JICA, in 2008 through a World Bank loan under the Diversified Farm Income and Market Development Project (DFIMDP), and in 2011 through a national government appropriated special fund under the Unified Enterprise GIS Project with the Department of Agriculture. We have the technical competency, we have the state-of-the-art equipment and facilities.
We have also updated our Soil Survey Manual to accommodate the latest developments in the USDA soil survey methods and soil profile description but we are yet to implement these changes because we are also yet to redesign our Soil Information System to accommodate the changes. Obviously, our soil survey and cartographic staff are yet to be trained on developments in various stages of digital soil mapping, except perhaps on the use of Global Positioning System (GPS) where this is already part of the field survey protocol.
We have also completed our manual on map standards and symbols for soil and water GIS. Standard procedures certainly differ for analog cartography with procurement and drafting of topographic maps to be used as base maps as against digital cartography where we just retrieve satellite imagery or the various thematic layers/shape files of pre-digitized base maps. With Mobile GIS, these satellite imageries and base maps can be loaded to personal digital assistance (e.g. hand-held tablets) for use in the field where field surveyors can put annotations and edit digital data that can be transmitted immediately to project management team in the central office through the internet. Map elements during map composition also differ in analog cartography and in digital cartography and we have already pre-defined as part of digital mapping standards the line specifications, markers, and annotations to imprint the so-called BSWM trademark for consistency of map outputs. Staff performance evaluation is being arranged with the Civil Service Commission to be dependent on number of metadata produced for various geographic datasets for us to be able to handle the complex and voluminous spatial data warehousing and retrieval operations that comes with digital soil mapping and cartography.
Nevertheless, digital soil mapping through the use of GIS/RS technological advances and computational advances such as application of geostatistical interpolation are being conducted but at experimental or at project level phase, not as a routine soil survey activity.
Why this Discrepancy?
First, is our organizational structure. We have one division handling the field soil survey activities, another division handling cartography, and another unit handling integrated soil resources information technology. Call it the relict of technological evolution in soil mapping through the decades.
Secondly, we are currently faced with hiring moratorium because of national government re-organization efforts. As the Bureau strives to keep up with the soil mapping developments, the hiring moratorium prevents us from filling up vacant positions with the so-called information age generation; leaving us with aging staff only familiar with traditional soil survey and traditional way of doing things, without understudies. With due respect of course to our technically competent soil surveyand cartographic staff, a saying goes that it is difficult to teach old dogs new tricks. Perhaps until we are re-organized and re-structured and the hiring moratorium is lifted, then we will be able to implement “digital” reforms and improvements in our routine soil survey activities.
Digital Soil Mapping Activities in BSWM
Digital soil mapping in BSWM is basically done at experimental or at project level phase. We can take a look at some of the past and current highlights.
- Geostatistical Interpolation of Soil Properties
- As early as 1993, we have attempted geostatistical interpolation using GEOEAS software and MS-DOS (286-based) computers for 4.25 km x 4.25 km spatial variability analysis and mapping of selected soil properties in reconnaissance soil surveys of Cavite province. By 2005, computer technology was far too advanced. We did a catenal variability of soil properties in our techno-demo farm in San Ildefonso, Bulacan;and the semi-variogram analysis and kriging for isarithmic mapping were done using ILWIS on Pentium IV at a fraction of time it took us to do the 1993 study. We could also play with the kriging results by overlaying topography as well as sampling points on the colored kriged results representing the level of interpolated soil property concentration. What a contrast to the MS-DOS computer output which was nothing more than isarithmic lines similar to topographic lines that took quite several minutes of computer-aided interpolation to produce.Today, with Core i5 and Core i7 processors and ArcGIS 10, ILWIS, and ENVI constituting our hardware and software inventory, respectively,kriging is but just a breeze and can be done as routine if it would be required by our soil surveyors. This of course could not be just done like that without collaboration with our Laboratory Services Division asgeostatistical interpolation efforts have implications on the number of soil samples to be analyzed in the laboratory. And this would have a corresponding implication on the national government appropriated funds for the Bureau considering escalating costs of laboratory chemicals.
- We used ALES developed by David G. Rossiter of Cornell University for land evaluation during those 286 IBM-PC platform days. Dr. Rossiter himself came to BSWM to introduce the software. Nowadays, we leave it to the staff. Some use very simple program, as simple as Microsoft Excel to generate the algorithm for FAO land evaluation framework for each of the soil mapping units, to more formal Soil Information System sub-set which we call the Soil Productivity Capability Classification System, to more sophisticated spatial data overlay using any commercial GIS softwares available in our Bureau. It is the staff level of competence to handle computer assisted evaluation that determines the level of agricultural land evaluation technology to be used. A number of our staff still rely on traditional manual rating.
- We have one staff, Dante Margate, who conducted a study using HYMAP airborne hyperspectralscanner and operated a GER3700 Spectroradiometer during the conduct of his masteral studies in ITC in 2000. He worked on an image acquired the year before, covering an area of 4 km breadth and 20 km length with spatial resolution varying from 5 to 10 m. The data was atmospherically corrected and converted into absolute reflectance using the ATCOR4 (2D) model. Georeferencing of the image, using only 3 bands, was carried out in ILWIS using 25 ground control points and applying affine transformation. A topographic map at a scale of 1:10,000 was used as a base map.
- Field spectral measurements were carried out to identify “desert-like” soil features such as surface accumulation of salts, calcium carbonate, and surface accumulation of gypsum materials. Reflectance spectra were obtained by comparing the radiance of the target with a reference panel (BaSO4 panel), made successively, and compared automatically to produce the reflectance measurements. Full profile descriptions, following the FAO-1990 Guidelines for Soil Profile Description were carried out in some selected observation points to understand the general soil distribution pattern in the study area, which was a semi-arid area of Tabernas, Almeria in southeastern part of Spain. Soil classification was carried out based on USDA Soil Taxonomy and World Reference Base for Soil Resources.
- The field identification of “desert-like” soil features were later confirmed by laboratory analyses. Spatial locations of these soil features were determined using Garmin 12XL GPS receiver. For post field work, spectral measurements of soil samples were also carried out at laboratory conditions using GER spectroradiometer to obtain the reflectance spectra. A 500 watts halogen lamp was used as light source positioned at about 15ozenital angle illuminating directly the sample.Spectral measurements at laboratory conditions were utilized to establish specific spectral response of each target parameter.
- The interpretation of aerial photographs was carried out according to geopedologic approach developed by Zinck in 1988 and the delineations were according to geologic units. A total of twenty five photographs at approximate scale of 1:20,000 covered the entire study area. An uncontrolled mosaic was produced and digitized in ILWIS as vector map. Goemetric corrections were done using 40 ground control points after converting into a raster format. The geocoded HYMAP image was used as the base map for the necessary corrections.
- Established reflectance curves, after matching with image spectra, were used to classify an atmospherically corrected airborne hyperspectral data of the study are. Spectral angle mapper, an algorithm which compares image spectra and individual spectra of specified targets, was applied. The spatial distribution of each “desert-like” soil feature was verified to the soil-landscape pattern of the area.
- A strong absorption feature at around 380-410 nm for all soil samples was selected and correlated with soil color parametes (hue, value, and chroma). Significant correlation existed between value and depth, value and area, value and asymmetry, and hue and width. The study asserted that based on the 33 soil samples analyzed, the mean organic matter content was only 0.7% and therefore, the masking effect of organic matter on reflectance of other soil constituents was minimal. Soils with relatively high silt content showed higher reflectance. Significant correlation coefficients were observed for electrical conductivity values with depth, width, and asymmetry at wavelength of 800-810 nm, implying that the depth of the absorption feature of this wavelength interval increased with the degree of salinity. The study also constructed a normalized difference vegetation index (NDVI) from the available bands of the image to show the extent of vegetation coverage. Using Spectral Angle Mapper, extensive areas were classified to be calcareous soils, around 40% of the total area. Saline soils occupied 17% while desert pavements covered 11%. Except for the overestimation of exposed gypsum and the occurrence of several pixels on unrelated units, the spatial distribution of “desert-like” soil surface features was found to be reasonable.
- We have another staff, Juliet Manguerra, who used spectral response of selected soils to assess their nutrient conditions. Three major agricultural soils – Alfisols, Inceptisols, and Ultisolsat varying levels of nitrogen, phosphorus, and potassium were collected and their spectral reflectance was measured in the laboratory using Ocean Optics S2000 Miniature Fiber Optic spectrometer at wavelength intervals of 400 to 850 nanometers. The spectral responses were analyzed using analysis of variance (ANOVA), spectral derivative analysis, and spectral resampling technique to sensors of selected satellite images. ANOVA was used to determine whether the soil orders and the different nutrient levels differ significantly in reflectance responses. Derivative analysis established the most responsive wavelength and the degree of differences in reflectance values among the samples. The Spectral Library Resampling menu of ENVI 4.0 was used to resample the spectral data file of the samples. The spectral data inputs were resampled to match the response of the instruments of Landsat TM, Landsat MSS, SPOT and ASTER satellite sensors. ENVI assumes critical sampling and uses the Gaussian model. Spectral resampling matched the measured spectral data using the spectral sensitivities or filter windows of the sensors of Landsat TM/MSS, ASTER and SPOT along the electromagnetic spectrum. The central wavelength absorption by soil order and nutrient conditions was derived for the respective bands of the satellite sensors. Spectral curve were plotted to the corresponding bands of Landsat TM/MSS, ASTER and SPOT. The ANOVA proved that there is a significant difference in the mean reflectance values of soil order and moisture content. Phosphorus-treated samples showed a remarkable difference in spectral response in all the soil orders as compared to those of nitrogen and potassium. Correspondingly, the increasing levels of nitrogen, phosphorus, and potassium by soil order demonstrated significant differences in the mean reflectance values. Spectral curves for the three soil orders and elemental nutrients have proven to be distinct along 400 – 850 nm of the electromagnetic spectrum.
- The derivative analysis established the most responsivewavelength and the extent of difference in reflectance values displayed the degree of separabilitybetween soil order and nutrient condition. The outcome of the resampling technique done has given comparable results with the spectrum derivative analysis; both have provided similarities in spectral patterns in all the soil types and nutrient conditions at specific wavelength in the visible region of the electromagnetic spectrum. The spectral data measured can be identified and the reflectance response varies in the different bands of SPOT, ASTER, and Landsat sensors. The established reflectance curves could support image classification analysis relative to soil nutrient mapping. Validation is crucial using other techniques. The possibility of development of spectral database specific to Philippine soils to support image analysis in soil fertility mapping is possible.
- We have a research project funded by the Australian Centre for International Agricultural Research (ACIAR) on Watershed Evaluation for Sustainable Use of Sloping Agricultural Lands in the Southern Philippines. The project has just completed the reconnaissance survey last December, 2011 and would commence three months of field survey work beginning February, 2012.
- The first objective is to improve the efficiency with which land attributes are mapped by using digital soil mapping techniques to produce more detailed and quantitative maps. Digital soil mapping (or, more broadly, digital land resource assessment) has two components. First is to develop methods of rapid soil measurement that allow quick and inexpensive (though less precise) measurements/ estimates of important soil attributes. This recognizes that, in the face of soil variability, the appropriate strategy is to make lots of inexpensive measurements rather than a few, precise and more expensive ones. Rapid soil measurement would allow measurements to be made at a far greater proportion of observation locations and increase the data density. The methods include the use of proximal sensors, such as visible/near infra-red and/or mid infra-red spectrometers, and pedotransfer functions to estimate soil attributes. These methods still require the use of conventional laboratory and field measurements on a proportion of sample locations, to provide local calibration.
- The second component of this objective is to improve spatial prediction of land attributes directly without the intermediate step of using soil type. This involves deriving statistical models of the relationships between land attributes (measured at point locations as discussed above) and spatial data covering the whole area of interest. Such data include digital elevation models and satellite imagery.
- Overall, the strategy is to adapt digital land resource assessment techniques to the conditions in Philippine upland watersheds to provide maps of land attributes with better utility and at lower cost. This will give a more complete picture of a watershed when combined with other data from rapid appraisal including land use and socio-economic information.
- This project will use spatial information to determine where technologies resulting from previous projects can best be located in the landscape, both by identifying vulnerable areas that require protection and their crop requirements. Spatial land suitability information will be used to enhance community and government agency decision making. This adds a new aspect to community participation and adds value to previous ACIAR projects in the Philippines.
- Digital soil mapping as a routine soil survey activity is definitely a long way to go in BSWM but certainly the only way to go considering the retirement of our soil surveyors and the moratorium on the hiring of understudies to continue their classical soil survey work. Once our reorganization plan is finally approved and we fill-up our vacant positions, only computer literates but classical soil survey illiterates would be on the job market. We are faced with very limited options with continually evolving technology and changing labor market. Nevertheless, we have the current equipment and the manpower to cope with the transition.
- There has been considerable development in digital soil mapping technologies and methodologiesin the recent decade and its application in traditional soil survey is expanding. For now, BSWM is into FAO-funded project on harmonizing its land degradation assessment with the global data set. We are also in the process of acquiring satellite imageries to cover the whole country for orthoimagery. Once processed, we intend to re-issue all our maps based on these orthoimages.
Summary and Conclusion
We are certainly eager to participate in this global consortium, not much for what we could get, but more so for what we could contribute. We recognize that we are a member of a global scientific community and we subscribe to what Jon Hempel was quoted during the GlobalSoilMap.net initial meeting in New York on 17 February 2009, “Soil attributes are critical inputs for ecosystem services. We need to provide a consistent set of data that is geographically continuous, scalable, and which includes uncertainty estimates”. We recognize the set of global standards which we need to conform with so that our outputs are consistent and in accordance with the specifications of and acceptable to the international soil science community.
We are grateful to FAO and to the organizing committee of this East Asia Node of the DigitalSoilMap.net for inviting us to participate in the launching of this partnership. We look forward to fruitful years ahead of us.
Back to Articles list Back to top
SRDC Bldg. Visayas Avenue corner Elliptical Road, Diliman, Quezon City, Philippines 1101
Tel No. (632) 332-9534 | Facsimile No. (632) 332-9534 | Email: email@example.com