Spatial analysis of groundwater quality data using geoR and mgcv R-package (ICMNS 2014)
Download as PPT, PDF1 like924 views
The document presents a first draft analysis of groundwater quality in the Bandung-Soreang groundwater basin, highlighting the degradation of groundwater conditions over time due to declining water levels and quality. It aims to identify the mixing processes between groundwater and surface water and to utilize water quality data for predicting groundwater flow. The findings suggest a close interaction between these water bodies and indicate potential influences from agricultural activities and geological controls on water quality.
Spatial analysis of groundwater quality data using geoR and mgcv R-package (ICMNS 2014)
- 1. First draft Spatial analysis of groundwater quality data using geoR and mgcv R-package Irawan, DE.1, Prabowo, K.1, and Akter, F.2, Vervoort, W.2 1 Faculty of Earth Sciences and Technology, Bandung Institute of Technology, Institut Teknologi Bandung, Jl. Ganesa No. 10, Bandung, 40132 2 Faculty of Agriculture and Environment, University of Sydney Biomedical Building, Australian Technology Park, NSW The 5th International Conference on Mathematics and Natural Sciences Institut Teknologi Bandung, 2-3 November 2014 1
- 2. Background First draft • The groundwater condition in Bandung-Soreang Groundwater Basin (BSGwB) has been degraded over time. Indications: • decline of water level and • decrease of water quality. • We need to check the occurring interactions between groundwater (especially shallow groundwater) and surface water. The 5th International Conference on Mathematics and Natural Sciences Institut Teknologi Bandung, 2-3 November 2014 2
- 3. Two main objectives • To identify: First draft • the intensity of mixing process in the groundwater and surface water. • if the water quality can be used to support groundwater flow prediction. The 5th International Conference on Mathematics and Natural Sciences Institut Teknologi Bandung, 2-3 November 2014 3
- 4. Regional setting • Bandung city area, Bandung-Soreang Groundwater Basin. • Groundwater-river water interaction have be mapped by Lubis (1997) and Rochman et al. (2014) J a v a s e a B a n d u n g W e s t J a v a Lembang fault Isolated Effluent Influent Cikapundung 0 4 E S N km W Maribaya Viaduct Dayeuhkolot The 5th International Conference on Mathematics and Natural Sciences First draft B a n d u n g C i k a p u n d u n g w a t e r s h e d C i t a r u m Institut Teknologi Bandung, 2-3 November 2014 4
- 5. Methods1 First draft • The dataset was composed of 295 points in total. • The data was measured in five years (1997, 1998, 2007, 2011, 2012). For each year we have 59 data: • 51 data from dug wells and springs. • 8 data from Cikapundung river. • All measurements were taken in July each year using We used handheld instruments (Hanna Instrument). • The dataset consists of 25 variables: • physical variables: x coordinate (x), y coordinate (y), elevation (elv), aquifer (aq), electroconductivity (ec), pH, hardness (hard), TDS, temperature, EH, discharge (Q) • cation variables: Ca, Mg, Fe, Mn, K, Na • anion variables: CO3, HCO3, CO2, Cl, SO4, NO2, NO3, SiO2 The 5th International Conference on Mathematics and Natural Sciences Institut Teknologi Bandung, 2-3 November 2014 5
- 6. Methods2 First draft • We introduce an statistical analysis to explore the dataset, consists of: spatial analysis and multiple regression (generalised additive model). • Open source and cross-platform software: • R3.1.1 (R Core Team, 2014) and Rstudio IDE0.98.1028 (RStudio Team, 2014). • Add-on packages of R: • graphical plotting using “lattice” (Sarkar, 2008), • data manipulation with “dplyr” (Wickham and Francois, 2014), • spatial analysis using “sp” and “geoR” (Roger S. Bivand, 2013), • Generalised additive model using “mgcv” (Wood, 2011), • principal component analysis using “pcaMethods” (Stacklies et al., 2007). The 5th International Conference on Mathematics and Natural Sciences Institut Teknologi Bandung, 2-3 November 2014 6
- 7. Methods3 First draft • Using R, we can make the spatial and regression model with straight forward codes. Here’s the example. • For spatial analysis using variogram model for electrical conductivity (EC), we used the following main code: > ECVariogramModel <- variog(EC, trend="1st”) > ECVariogramModel <- variog(EC, max.dist=25000, uvec=seq(0,25000,by=5000)) The 5th International Conference on Mathematics and Natural Sciences Institut Teknologi Bandung, 2-3 November 2014 7
- 8. Methods4 First draft • For regression model for electrical conductivity (EC) in the function of (x,y) coordinates, elevation, and aquifer type, we used the following main code: LinearModel <- gam(EC ~ te(x, y, k=k1, bs=bsm) + s(Elevation, k=k1, bs=bsm) + (AquiferType), data=BandungBasin)) The 5th International Conference on Mathematics and Natural Sciences Institut Teknologi Bandung, 2-3 November 2014 8
- 9. Results1: Mixing process First draft • All plots suggest the possible mixing process between groundwater and river water. • Some parameters demonstrate: – decreasing pattern towards the river: EC, CO3, CO2, SO4, and NO2. – increasing pattern: NO3, Fe, Mn, and Na. The 5th International Conference on Mathematics and Natural Sciences Institut Teknologi Bandung, 2-3 November 2014 9
- 10. Results2: Mixing process First draft • The water quality supports effluent stream flow (groundwater flows from aquifer to the river) from the previous water flow mapping. • The NO2 and NO3 reflect the influence of agricultural activities in the upstream part. • Fe, Mn, and Na are traces of geological control which composed mostly of volcanic rocks. • The other significant elements (EC, CO3, CO2, SO4) are possibly coming from the general domestic activities. The 5th International Conference on Mathematics and Natural Sciences Institut Teknologi Bandung, 2-3 November 2014 10
- 11. Results3: Mixing process • Both analyses draw a possible mixing process due to the close interaction between both water bodies. • By plotting the principal component, we can still separate the river water samples from groundwater cloud. The 5th International Conference on Mathematics and Natural Sciences First draft Institut Teknologi Bandung, 2-3 November 2014 11
- 12. Results2: Proposed water quality model RW flow High NO3 GW flow: Add NO3 GW flow Shallow GW flow Cl, SO4 from deeper GW potentially drawn up due to heavy pumping Decreasing EC, Na, Fe, NO2, SO4, Mn, CO3, CO2 due to river dilution The 5th International Conference on Mathematics and Natural Sciences First draft GW signature (from lithology): Ca, Mg, Na, K aren’t significant -> somehow masked? Institut Teknologi Bandung, 2-3 November 2014 12
- 13. Conclusions First draft • Both analyses draw a possible mixing process due to the close interaction between both water bodies. However, by plotting the principal component, we can still separate the river water samples from groundwater cloud. • Considering the number of data in the training dataset, we can use this plot to predict the water character to another water quality dataset from the area The 5th International Conference on Mathematics and Natural Sciences Institut Teknologi Bandung, 2-3 November 2014 13
- 14. Acknowledgement First draft • We wish to express our gratitude to Prof.Dr. Eddy A. Subroto as the Dean of FEST and Prof.Dr. Lambok Hutasoit as Head of Applied Geology Group for their continuous support for this collaborative research between ITB and University of Sydney. Main references: 1.BPLHD, Annual Environmental Report of West Java Province (Badan Lingkungan Hidup Daerah Provinsi Jawa Barat, 2010). 2.Cavazza, S., Pagliara, S., "Groundwater and surface water interactions" , EOLSS Encycl. Life Support Syst. 1 (2000). 3.Diggle, P.J. and Ribeiro Jr, P.J. Model Based Geostatistics (Springer, New York, 2007). 4.Geyh, Isotopic Hydrological Study in Bandung Basin, Indonesia (CTA-PU Project Report, 1990). 5.Hutasoit, LM, “Simulasi Numerik CAT Bandung-Soreang” Jurnal Geologi Indonesia (1)1 (2008). 6.Koesoemadinata dan Hartono, “Sedimentasi dan Stratigrafi Daerah Bandung”, Prosiding PIT IAGI (Ikatan Ahli Geologi Indonesia, 1981). 7.Matahelumasi, B. and Wahyudin, Penelitian Hidrogeologi Daerah Imbuhan Air Tanah dengan Metode Isotop dan Hidrokimia, di Cekungan Air Tanah Bandung Soreang, Provinsi Jawa Barat (Tahap II) (Pusat Lingkungan Hidup dan Geologi, Badan Geologi, 2009). 8.Priowirjanto, G and Marsudi, “Fluktuasi Muka Air Tanah di Cekungan Bandung”, Prosiding Seminar Sehari Air Tanah Cekungan Bandung (Perhimpunan Ahli Airtanah Indonesia, 1995). 9.R Core Team, R: A Language and Environment for Statistical Computing (R Foundation for Statistical Computing, Vienna, Austria, 2014). 10.Roger S. Bivand, V.G.-R., Edzer Pebesma, Applied spatial data analysis with R, Second edition (Springer, New York, 2013). 11. RStudio Team, RStudio: An Integrated Development Environment to R (RStudio, Vienna, Austria, 2014). 12.Sarkar, D., Lattice: Multivariate Data Visualization with R (Springer, New York, 2008). 13.Silitonga, Peta Geologi Lembar Bandung, Jawa Barat, Skala 1:100.000 (Pusat Penelitian dan Pengembangan Geologi, 1973). 14.Sudjatmiko, Peta Geologi Lembar Cianjur, skala 1:100.000 (Pusat Penelitian dan Pengembangan Geologi, 1972). 15.Sunarwan, B., “Analisis Unit Hidrostratigrafi CAT Bandung-Soreang” (Disertasi, Program Studi Teknik Geologi ITB, 2014). 16.Sutrisno, Peta Hidrogeologi Lembar Bandung, Jawa Barat, Skala 1:100.000 (Pusat Geologi Tata Lingkungan, 1983). 17.Wickham, H., Francois, R., dplyr: A Grammar of Data Manipulation (Springer, New York, 2014). 18.Wood, S.N., "Fast stable restricted maximum likelihood and marginal likelihood estimation of semiparametric generalized linear models", Jour. of the Royal Statistical Soc (B) 73(1) (2011):3-36. The 5th International Conference on Mathematics and Natural Sciences Institut Teknologi Bandung, 2-3 November 2014 14
- 15. First draft Spatial analysis of groundwater quality data using geoR and mgcv R-package Irawan, DE., Prabowo, K., and Akter, F., Vervoort, W Thank you for your attention Further questions and comments should be addressed to: Dr. Dasapta Erwin Irawan : erwin@gc.itb.ac.id Twitter : @dasaptaerwin Due to the limited time to present and pages to write, the resources from this project will be made available online by the end of this month at: •ResearchGate: https://www.researchgate.net/profile/Dasapta_Irawan •Academic.edu: https://itb.academia.edu/dasaptaerwin To the extent possible under law, the author of this document have waived all copyright and related or neighboring rights to this work. This work is published from Sydney, Australia. The 5th International Conference on Mathematics and Natural Sciences Institut Teknologi Bandung, 2-3 November 2014 15