East Kalimantan Programme; MAHAKAM

WP2: Quantifying sediment fluxes in the Mahakam Delta using remote sensing

Research group
Project leaders:
Dr. Zoltán Vekerdy
Dr. Orbita Roswintiarti

Researcher:
Syarif Budhiman

Other participants:
Prof. Dr. Z. (Bob) Su
Dr. Chris Mannaerts
Dr. Jan de Leeuw
Dr. Valentyn Tolpekin
Dr. Marieke Eleveld

3. Summary of the project
Sedimentation processes in the Mahakam Delta are governed by the discharge and sediment transport of the Mahakam River, the tide and the coastal currents. Mangroves set a natural boundary condition for theses complex processes, and human activities (especially by converting mangroves into fish ponds) express the socio-economic pressure on the delta. Remote sensing allows monitoring sedimentation processes in time and space as was demonstrated in the pilot phase of the East Kalimantan Programme.
The main objective is to quantify sediment fluxes by fluvial and tidal processes in the Mahakam delta over the last 50 years using data from aerial photographs, optical and microwave remote sensing techniques, in combination with sediment transport modelling as well as field data on sediment concentration in the delta distributaries. In this way, a reconstruction of delta development in the last fifty years can be obtained, and future trends may be predicted.
Only optical remote sensing allows the direct study of suspended sediments (TSM) in the water, but this tool fails in case of cloud cover. Microwaves penetrate clouds, but do not penetrate the water body. Therefore, an innovative synergy of the different remote sensing image types with ancillary data is planned for the reconstruction of the history of delta development.
Imaging spectrometry using bio-optical models is feasible for mapping TSM in the cloud-free periods. Spectra of water-leaving radiance are of importance to test bio-optical model closure. Therefore field measurements and laboratory analyses are needed. Based on the determination of the inherent optical properties of the water constituents a bio-optical model will be calibrated. Model inversion techniques will be used to obtain water quality parameters from remote sensing images in order to monitor water quality.
In the cloudy wet season, when most of the sediment is produced from the basin, indicators of sedimentation and the dynamics of the branches will be analysed based on microwave images, in combination with the results of sediment yield studies by in-situ measurements and sediment transport modelling. Rating of sediment transport as a function of water stage, i.e. water cover defined from radar images, will allow the assessment of sediment load in the delta even under cloudy conditions. The results will allow the analysis of changes attributable both to natural processes (e.g. climate change) and human interactions.

4. Detailed description of the project
a. Scientific Background
The Mahakam delta is unique in its virtual separation of the pathways of fluvial and tidal sediment fluxes (see description WP1). However, sedimentation processes in the Mahakam Delta show high variability in space and time (Storms et al. 2005), and predicting its future development under various scenarios of natural and anthropogenic controls requires a thorough understanding of this spatial and temporal variability. A preliminary bio-optical model based on Remote Sensing imagery and field data on Total Suspended Matter concentrations in the delta waters collected during the first pilot phase of the EKP programme showed that fluvial distributaries and tidal channels react virtually independently to changes in river discharge and tidal amplitude (Figures 1 and 2). At the other hand, the impact of the recent conversion of mangrove swamps to shrimp ponds in the delta could not yet be demonstrated (Budhiman, 2004). The present project aims to (1) refine this model, (2) device new methods to obtain sediment concentrations from radar imagery in the mousson season when optical remote sensing is impeded by cloud cover, and (3) to extend the model output backwards in time using older data. In this way we want to obtain a quantitative model of sediment fluxes for the last 50 years, which will be incorporated into the 3DSEDFLUX/DELFT3D model developed in WP1.

Modelled Total Suspended Matter concentration at high river discharge and low tidal amplitude: mainly fluvial distributaries are active (Landsat TM), Budhiman, 2004

Modelled Total Suspended matter concentration at low river discharge and high tidal amplitude: mainly tidal channels are active (SPOT HRV), Budhiman, 2004

(1) In the wavelength range of 0.4-0.7µm, light penetrates into the natural water bodies and a complex interaction takes place (absorption, scattering, emission and reflection). The spectral characteristics of the backscattered, water-leaving energy depend on the quality constituents of the water body. In optically deep water bodies, if the optical properties of the water and its constituents (as a function of their concentrations) are known then their contribution to the water column colour can be discriminated and their concentrations can be quantified (Dekker et al. 2001) by inverting the optical models, which describe these dependencies. For satellite RS, the inversion of such models needs also the explicit modelling of the atmospheric effects on the detected energy.
The relation between sediment concentration and light reflection as found in the field can be translated into empirical and semi-analytical algorithms for mapping suspended matter (Kirk 1983). Such procedures were developed for TSM concentration mapping in the Mahakam Region (Ambarwulan et al. 2004; Budhiman et al. 2004). These regression models related TSM values defined by water sampling to RS radiances (Fig 1, 2).
As the pilot phase of the East Kalimantan programme showed (Budhiman 2004) uncertainties are involved in TSM mapping of the coastal waters around the Mahakam Delta, when model parameters determined in another geographic region are used. Similar experience was reported by Turdukulov et al. (2003) about experiments in Europe. In coastal areas, the optical depth of the water and the substrate influences the appearance of the water body on the images (Phinn et al. 2004). Inevitably, these facts underlined the importance of the accurate determination of the inherent optical properties (IOPs) of the water. Understanding and removing the effect of atmospheric disturbances on the images requires in-situ measured reference spectra of quasi-invariant objects. Spectra of water-leaving radiance (taken simultaneously with IOP measurements) are also of importance to test bio-optical model closure (Hakvoort et al. 2002).
Further research is needed to define the spectral characteristics of the water-leaving radiance of river branches with different fluvial and tidal influence as well as of coastal waters. Furthermore, it is necessary to define the inherent optical properties of the water and the suspended sediments in laboratory, to allow the setup and inversion of bio-optical models for remote sensing of TSM with no (or very little) synchronous field measurements.

Preliminary model to derive sediment concentrations from field and optical RS data (Budhiman, 2004)

(2) Radiation of optical wavelengths does not penetrate clouds, so satellite RS of water quality is not possible in large part of the year in the wet tropical conditions of East Kalimantan, when the most of the discharge and sediment load arrives to the delta.
The applicability of radar images for mapping water-covered areas under cloud cover is widely recognised in the literature, e.g., (Hess et al. 1990), but with radar, it is not possible to get direct information about the water quality constituents of the water body. Microwaves penetrate clouds, but interact with the water differently from the optical waves: no penetration takes place, so no water quality information can be deduced from the backscattered energy. Lower frequency radar images (e.g. Lband) penetrate vegetation canopies; relatively little scattering takes place from the canopy and the underlying inundation can be mapped (Townsend et al. 1998). This information allows the estimate of discharges even in much braided river channels (Smith et al. 1996). Radar images of higher frequency, e.g. (multi-polarized) X-band images provide information about the canopy structure, allowing the identification of vegetation classes in mangroves.

b. Specific Objective(s)
The main objective of the research is to reconstruct sediment fluxes and delta development in the last decades using synergetic RS methods, field data on sediment concentrations and field spectral measurements, and predict future trends. The project focuses on three specific objectives:

Validate/upgrade sediment models with optical RS (TSM concentration mapping) of moderate to high spatial resolution images using multi-spectral images (e.g., SPOT) and hyper-spectral images (e.g. Hyperion and if possible: CASI data). Main questions are:

Under which conditions (wave height, water depth, disturbances of light by water surface, etc.) do the algorithms work and what is the accuracy of the algorithms?
How the optical properties of different suspended sediment fractions (separated by grain size, geological origin, etc.) effect the specific optical properties of the turbid waters?
Is there temporal variability in sediment reflectance (e.g. due to variability in grain size distribution), and if so, what are the consequences on the interpretation of sediment transport from optical satellite images?

Estimate TSM load during cloudy periods using information deduced from radar images and applying sediment rating curves and/or sediment transport models. Main questions are:

How to derive water extent from radar data (L-band), and what is the mapping accuracy?
How can the relation between the flooded area and the water/sediment discharge be deduced using radar images?
Can vegetation structure data derived from X-band images be related to resistance to flow, hence to sedimentation during high water periods?

Retrospective monitoring of sediment concentrations in older RS imagery of the Mahakam Delta with RS Imagery (e.g. optical and radar images, including aerial photos - if available) and old maps.

How did the delta develop in the last decades, and how can active and inactive delta branches be differentiated in the different RS images?
How quantitative data on sediment pathways and concentrations can be derived from older imagery (especially aerial photographs) etc.)?
How can errors e.g. in imaging geometry, canopy cover, mixed pixels, different spatial and spectral resolutions, be minimized?

c. Workplan
1. Preparatory phase
Selection of imagery types, key areas and key periods for field measurements during satellite overpass; analysis of older imagery. Spectral and geometric pre-processing; data fusion to derive the river bank and coastline changes; image classification. Due to the inhomogeneous data sources, postclassification change detection methods will be applied, but the efficiency of these methods will be maximized with proper pre-processing techniques (Petit et al. 2001). The first fieldwork will be partly spent on identifying the main classes/features to support the image classification.

2. Field sampling
During the sampling cruises stratification of water masses will be investigated using a Seabird CTD (Conductivity Temperature Depth) probe, which will be sampled synchronously with the Parametric Echosounder observations (see WP1). The survey will be planned to be simultaneous with satellite data acquisition. The CTD casts will be taken using a winch on the research vessel. Vertical profiles of salinity will be obtained using the upcasts only, covering the entire water column except for the bottom 50 cm. A Seapoint Optical Backscatter sensor (OBS) will be attached to the CTD probe. For the OBS measurements, only the downcasts will be used, as the CTD probe could create some sediment clouds when touching the seabed. For calibration purposes, a Niskin bottle will be attached parallel to the CTD to collect in situ water samples.
Water samples will be taken at a depth up to 30 cm into 500 ml plastic bottles (3 bottles for each location) and stored in a freezer for keeping the organic matter constatnt. Two bottle samples will be analysed for TSM and Chl content respectively in the Laboratory of Limnology at IPB (Bogor Agricultural University). The suspended sediment content of the water samples will be measured by vacuum filtration of a fixed amount of water on pre-weighed polycarbonate filters with a pore size of 0.4 Am. After filtration, the filters will be cleaned with nano-pure water to remove salts, washed with alcohol, and dried and weighed. The third series of bottles will be analysed at the Free University of Amsterdam for Inherent Optical Properties determination, including beam attenuation, CDOM (“aquatic humus”) absorption and seston absorption.

3. Field spectrometry measurements
Field spectrometry will be carried out using a GER 3700 and ASD field spectrometers to obtain the subsurface irradiance reflectance R(0), which is the ratio of upwelling radiance and downwelling radiance just beneath the surface. These data will be compared with the IOP (Inherent Optical Properties) determined in the laboratory from water samples taken at the same spot.

4. Estimate TSM load during cloudy periods using information deduced from radar images
In determining the trends, the work is to be based on a strong cooperation with project partners working on the sediment yield of the watershed. Most sediment transport takes place in the high-water periods (especially in the rising phase), i.e. in the wet seasons, when most of the sediment is produced from the basin (Renschler et al. 1999). Methods for suspended sediment estimation directly from optical RS images are not applicable in these periods. Thus indirect methods will be used, those we plan to base on radar images (Hessner et al. 2001).
Indicators of erosion and sedimentation (the dynamics of the branches) based on microwave images will be statistically analysed in combination with the results of the sediment yield studies based on in situ measurements. Vegetation traps sediments, so long-term analysis of the mangroves can provide an indication of the delta processes, as was shown using (optical) Spot images by (Dutrieux et al. 1990). Based on the analysis of multi-polarized X-band backscatter from the canopy (TerraSAR-X) mangrove structure will be correlated to sedimentation measured in the field. Testing of Envisat-ASAR images for canopy structure mapping is also planned. Techniques for speckle reduction of radar images will be used.
Lower frequency (e.g. L-band signals) have a deeper penetration in the canopy. PALSAR images from the Japanese ALOS satellite will be used for mapping water-covered areas under clouds. The high spatial resolution of these new generation radar images will allow a detailed mapping of the land cover. Strict work plan cannot be presented here, because data availability is not secured yet. LAPAN has an agreement with the Japanese Aerospace Exploration Agency (JAXA) to develop scientific applications for the PALSAR images. Corner reflectors have been installed in the Mahakam Region for the geometric correction of Envisat-ASAR images. ITC is a member of the TerraSAR-X user community. These facts allow the here-proposed project to get access to images after the satellites are operational.

5. Upgrade sediment models with optical RS, radar and applying sediment rating curves and/or sediment transport models, and extension back in time
Based on field sampling and laboratory determination of the IOPs of the water constituents a bio-optical model will be calibrated that is able to correctly simulate observed spectra using measured concentrations. Model inversion techniques will be used to obtain water quality parameters from RS images in order to monitor water quality. Sediment rating curves will be created using the images and field data. We will try to extend the model to earlier dates on the basis of the synergy of archive RS images and ancillary data, based on the antecedent related efforts of LAPAN. We plan to use topographic maps (earliest 1930s), aerial photographs (earliest ?) Landsat MSS data (starting from the beginning of the1970s), Landsat TM data (1980’s and 1990’s), SPOT images (from the 1980’s), ASTER data (from 2000), and Landsat ETM data (2000’s). Radar data from archives (ERS1, ERS2, Radarsat) will be used for mapping the delta under cloud cover. We will use the archives and data services of the ground receiving station of LAPAN at Parepare (South Sulawesi). In this way we hope to improve our understanding of sediment dynamics at least 50 years back.

d. Scientific Relevance
We hope to develop a physically based, generic optical model, which makes the TSM estimates more accurate and more independent from ground measurements.

The planned synergy of optical and radar images will provide new information about mangrove structure and delta morphology. This will result in new experience in the application of the images of the latest sensor generation and a deeper insight into the processes in the Mahakam Delta, serving the needs of coastal zone managers (Eleveld et al. 2003) here and in similar coastal areas under the tropics.

d. Scientific Relevance
We hope to develop a physically based, generic optical model, which makes the TSM estimates more accurate and more independent from ground measurements.

The planned synergy of optical and radar images will provide new information about mangrove structure and delta morphology. This will result in new experience in the application of the images of the latest sensor generation and a deeper insight into the processes in the Mahakam Delta, serving the needs of coastal zone managers (Eleveld et al. 2003) here and in similar coastal areas under the tropics.

5. Participation in a graduate School ('onderzoeksschool')
The project will be carried out in cooperation with:
Production Ecology and Resource Conservation (PERC)
Wageningen University
Haarweg 333
6709 RZ Wageningen
The Netherlands

Boussinesq Centre for Hydrology (Utrecht University, Delft University of Technology, Free University of Amsterdam, Wageningen University, UNESCO-IHE and ITC)

Twente Centre for Water Systems and Government

6. Scientific performance of members of the research group(s)

Kardeván, P., Z. Vekerdy, L. Róth, S. Sommer, T. Kemper, G. Jordán, J. Tamás, I. Pechmann, E. Kovács, H. Hargitai and F. László (2003). Outline of scientific aims and data processing status of the first Hungarian hyperspectral data acquisition flight campaign, HySens 2002 Hungary. 3rd EARSeL Workshop on Imaging Spectrometry, Herssching, Germany, EARSeL: 324-332.
Meijerink, A. M. J. and Z. Vekerdy (2003). Satellite images for the monitoring of wetlands and assessing their water budgets. In: Geoinformatics for tropical ecosystems. P. S. Roy. Dehra Dun, India, Bishen Singh Mahendra Pal Singh: 513-538.
Meijerink, A. M. J., A. S. M. Gieske, Z. Vekerdy (2005). "Surface energy balance using satellite data for the water balance of a traditional irrigation - wetland system in SW Iran." Irrigation and Drainage Systems 19(1): 89-105.
Su, Z., T. Schmugge, W.P. Kustas, W.J. Massman, 2001, An evaluation of two models for estimation of the roughness height for heat transfer between the land surface and the atmosphere, Journal of Applied Meteorology, 40(11), 1933-1951.
Wen, J., Z. Su, Y. Ma, 2003, Determination of Land Surface Temperature and Soil Moisture from TRMM/TMI Remote Sensing Data, Journal of Geophysical Research, 108(D2), 10.1029/2002JD002176.
Budhiman, S. (2004). Mapping TSM concentrations from multisensor satellite images in turbid tropical coastal waters of Mahakam Delta, Indonesia. M.Sc. thesis, Enschede, International Institute for Geo-information Science & Earth Observation (ITC): 82
Budhiman, S., T. Hobma and Z. Vekerdy (2004). Remote sensing for mapping TSM concentration in Mahakam Delta: an analytical approach. 13th OMISAR Workshop Validation and Application of Satellite Data for Marine Resources Conservation. 5 – 9 October 2004 Kuta, Bali, Indonesia.
Budhiman, S., R. Dewanti and C. Kusmana (2002). Application of Lansat TM and GIS for inventorying the degradation of mangrove forest in East Kalimantan. 6th Pan Ocean Remote Sensing Conference (PORSEC), 3-6 September 2002, Bali, Indonesia.
Harsanugraha, W.K., B.S. Tejasukmana and S. Budhiman (2000). Analisis Potensi Mangrove dan Tambak di Pulau Bali Menggunakan Data Landsat-TM, Majalah LAPAN edisi Penginderaan Jauh No. 01 Vol. 02 Maret 2000, LAPAN
Eleveld, M.A., Pasterkamp, R. & Woerd, H.J., van der (2004). A survey of total suspended matter in the southern North Sea based on the 2001 SeaWiFS data. EARSeL eProceedings, 3(2), 166-178. CD & URL: http://las.physik.uni-oldenburg.de/eProceedings
Eleveld, M. A., S. T. Block, et al. (2000). "Deriving relief of a coastal landscape with aerial video data." International Journal of Remote Sensing 21(1): 189-195
Bagheri, S., Peters, S.W.M., Yu, T: (2005): Retrieval of marine water constituents from AVIRIS data in the Hudson/Raritan Estuary, Int. J. rem. Sens.,Vol. 26(18), 20 September 2005, 4013-4027.

7. Literature references

Ambarwulan, W. and T. W. Hobma (2004). Bio-optical model for mapping spatial distribution of total suspended matter from satellite imagery. 3rd FIG Regional Conference, Jakarta, Indonesia
Budhiman, S. (2004). Mapping TSM concentrations from multisensor satellite images in turbid tropical coastal waters of Mahakam Delta, Indonesia. Enschede, International Institute for Geoinformation Science & Earth Observation (ITC): 82
Budhiman, S., T. W. Hobma and Z. Vekerdy (2004). Remote sensing for Mapping TSM concentration in Mahakam Delta: an analytical approach. 13th OMISAR Workshop on Validation and Application of Satellite Data for Marine Resources Conservation, Kuta, Bali, Indonesia
Dekker, A. G., V. E. Brando, J. M. Anstee, N. Pinnel, T. Kutser, E. J. Hoogenboom, S. W. M. Peters, R. Pasterkamp, R. Vos, C. Olbert and T. J. M. Malthus (2001). Imaging spectrometry of water. Imaging spectrometry: Basic principles and prospective applications. F. van der Meer and S. M. de Jong. Dordrecht NL, Kluwer Academic Publishers. 4: 403
Dutrieux, E. (1991). "Study of the ecological functioning of the Mahakam Delta (East Kalimantan, Indonesia)." Estuarine, Coastal and Shelf Science 32(4): 415-420
Dutrieux, E., J. Denis and J. Populus (1990). "Application of SPOT data to a base-line ecological study of the Mahakam delta mangroves (East Kalimantan, Indonesia)." Oceanologica acta 13(3): 317- 326
Eleveld, M. A., W. B. H. Schrimpf and A. G. Siegert (2003). "User requirements and information definition for a virtual coastal and marine data warehouse." Ocean & Coastal Management 46(6-7): 487-505
Hakvoort, H., J. d. Haan, R. Jordans, R. Vos, S. Peters and M. Rijkeboer (2002). "Towards airborne remote sensing of water quality in The Netherlands — validation and error analysis." ISPRS Journal of Photogrammetry and Remote Sensing 57: 171-183
Hess, L. L., J. M. Melack and D. S. Simonett (1990). "Radar detection of flooding beneath the forest canopy: a review." International Journal of Remote Sensing 11(7): 1313-1325
Hessner, K., A. Rubino, P. Brandt and W. Alpers (2001). "The Rhine outflow plume studied by the analysis of synthetic aperture radar data and numerical simulations." Journal of Physical Oceanography 31(10): 3030-3044
Kirk, J. T. O. (1983). Light and photosynthesis in aqatic ecosystems. Camberra, Australia, Cambridge University Press
Petit, C. C. and E. F. Lambin (2001). "Integration of multi-source remote sensing data for land cover change detection." International Journal of Geographical Information Science 15(8): 785-803
Phinn, S. R., A. G. Dekker, V. E. Brando and C. M. Roelfsema (2004). "Mapping water quality and substrate cover in optically complex coastal and reef waters: an integrated approach." Marine Pollution Bulletin In Press, Corrected Proof
Renschler, C. S., C. M. Mannaerts and B. Diekkrüger (1999). "Evaluating spatial and temporal variability in soil erosion risk-rainfall erosivity and soil loss ratios in Andalusia, Spain." Catena 34: 209-225
Smith, L. C., B. L. Isacks, A. L. Bloom and A. B. Murray (1996). "Estimation of discharge from three braided rivers using synthetic aperture radar satellite imagery: Potential application to ungaged basins." Water Resources Research 32(7): 2021–2034
Storms, J. E. A., R. M. Hoogendoorn, R. A. C. Dam, A. J. F. Hoitink and S. B. Kroonenberg (2005). "Late-Holocene evolution of the Mahakam delta, East Kalimantan, Indonesia." Sedimentary Geology 180(3-4): 149-166
Townsend, P. A. and S. J. Walsh (1998). "Modeling floodplain inundation using an integrated GIS with radar and optical remote sensing." Geomorphology 21: 295-312
Turdukulov, U. and Z. Vekerdy (2003). Determination of water quality parameters using imaging spectrometry (case study for Sajó flood plain, Hungary). 3rd EARSeL Workshop on Imaging Spectrometry, Herrsching, Germany, EARSeL: 454-463. Internet: