GIS BASED WATER BALANCE STUDY FOR ESTIMATION OF RUNOFF IN A SMALL RIVER WATERSHED

INTRODUCTION Water resources in Madhya Pradesh are mainly exploited for domestic, industrial water supply and development of irrigation scheme thus state is facing the problem of water scarcity. In such circumstances, understanding the complex system of hydrological processes and the water availability in the river basins is needed for the sustainable water resources management and development of an area which can be achieved by calculating the water balance in a distributed scale considering the Spatial variation due to distributed land use, soil texture, topography, and hydro-meteorological conditions and then representing the same through a hydrological model. Water balance is defined as the net change in water, by considering all the inflows and out flows of a hydrologic system. The development of various water resources management and modelling techniques were emphasized during the latter half of the last century and the main thrust was on improvement in the model efficiency. Subrahmanyam (1983) presented the principles and procedures of drought climatology employing the water balance concepts using Thornthwaite's index of aridity. Pimenta (1999) used the Thornthwaite and Mather s approach to compute the number of months of water stress conditions and amounts of water involved in the different components of the hydrologic cycle with a monthly time step, using GIS for Portugal. Kumar (2001) computed the normal water balance components such as actual evapotranspiration (AET), water deficit, & water surplus of Krishnai basin using the procedure of Thornthwaite (1948). Singh, et al. (2004) conducted a water balance study in Nana Kosi watershed, in the district of Almora (Uttranchal) using the Thornthwaite and Mather (TM) model with the help of Remote Sensing and GIS and besides showing the seasonal pattern of precipitation, actual evapotranspiration (AET), potential evapotranspiration Journal of Indian Water Resources Society, Vol 35, No.3, July, 2015 GIS BASED WATER BALANCE STUDY FOR ESTIMATION OF RUNOFF IN A SMALL RIVER WATERSHED Priyanka Gupta 1, T.R. Nayak 2 and M.K. Choudhary 3 ABSTRACT With the increasing demand and decreasing supplies of water, the water crisis is growing critically. Thus a shortage of water in future is inevitable unless we look for some technical ways of using water more efficiently. It is must to have an idea about the water availability and water requirements. So, water balance studies on river basins are important because it provides quantitative information on water availability and water requirements. In the present study an attempt has been made to compute the water balance of Bina river watershed, a major tributary of Betwa river in Madhya Pradesh using the Thornthwaite and Mather s model. The land use, soil texture, and other watershed parameters required as input to the model have been generated with the help of Remote Sensing and GIS techniques. The observed values of rainfall and runoff for the year 2007-2010 have been utilized for evaluation of the model. The results show that the yearly potential evapotranspiration in the watershed is 1229.31 mm. The actual evapotranspiration depends on the available soil moisture, viz. the duration and quantity of rainfall. The annual runoff in the basin is estimated to be 45.5% of the annual rainfall, which is high due to rocky & hilly terrain. The study reveals that the streams are generally dry in the months of November to June. Groundwater recharge (Soil moisture storage) takes place during July, however July, August, September & October months are the period of water surplus. Key words: PET; AET; Water Balance; Water Surplus; Water Deficit; Remote Sensing; GIS. 1. M.Tech. Scholar, National Institute of Technology, Bhopal 2. Scientist E & Head, National Institute of Hydrology, Bhopal 3. Associate Professor, National Institute of Technology, Bhopal Email: priyanka009.gupta@gmail.com, trnnca@gmail.com, mkchoudhary67@yahoo.co.in Manuscript No.: 1399 (PET) and runoff, periods of moisture deficit and soil moisture recharge were also indicated. Killingtveit (2004) through his study found that the water balance for the two catchments at Svalbard has an average residual term close to zero, but there were large errors in individual years with positive and negative deviations, the reason for this was that the individual terms in the water balance were not known well enough. Victoria, et al. (2006) simulated the monthly water balance for the Ji-Parana river basin, in the Western Amazonian state of Rondonia using Thornthwaite Mather climatological model and compared the observed discharge data with the modeled results, which indicated an under estimation of basin ET and an excess water surplus. It was suggested that these results could be due to underestimation of PET, rooting depth estimates, or a combination of both. But an increase in PET improved modeled results. Jasrotia, et al. (2009) assessed the water balance of Devak Rui watershed of Kandi region in Jammu district using the Thornthwaite and Mather (TM) model with the help of Remote Sensing and GIS and a runoff potential map was generated. Jenifa, et al. (2010) observed that traditional approach of calculating the water balance using a spatially and temporally lumped scale does not give very accurate estimate of the water volume in a hydrological component so a spatially semi distributed, GIS based hydrological model was developed on a sub watershed scale using mean monthly hydro-meteorological data Amaravathi River Basin and surface runoff and evapotranspiration were estimated using land use, soil texture, topography, and hydrometeorological parameters. Roy, et al. (2012) attempted to find a solution of water shortage problem in Paramount farm in Southern San Joaquin Valley, California. The resources were estimated by a water balance assessment approach using the Thornthwaite and Mather (TM) model. Jain (2012) on the basis of the estimates of range of ET for the areas which are hydro-meteorologically similar to India, found that ET for India is underestimated, and this is because the trans-boundary basins bring in large quantities of flow that has been generated beyond the borders by surface and sub-surface routes. Karsili (2013) presented a water balance model based on the Thornthwaite s methodology of the Mediterranean Region to understand the impacts of climate changes using the ArcGIS Model Builder. 26

In the present study efforts have been made to analyze meteorological and hydrological aspects of water availability in Bina river watershed in Madhya Pradesh, using the Thornthwaite and Mather model with the help of Remote Sensing and GIS techniques. The monthly water balance has been computed by considering the precipitation, evapotranspiration and soil moisture storage of the watershed. Subsequently, the computed monthly runoff has been compared with the observed monthly runoff. STUDY AREA The Bina river is a major tributary to the Betwa river in Yamuna basin, which originates from the Vindhyan hills in Raisen district and major part of the Bina river falls in Sagar district in Bundelkhand region in Madhya Pradesh. The study has been conducted in the Bina river watershed up to Rahatgarh Gauge & Discharge site (Fig. 1) in order to validate the model with the observed flow. The geographical area of the Bina river watershed up to Rahatgarh is about 1139.03 Km 2. The Bina river in study area traverses through the undulating topography mostly covered with open forests and scrubs and the gently sloping foot hills and plain area covered with agriculture landuse. The study area is located between 24 10 to 24 42 N latitudes and 78 09 to 78 23 E longitudes. Four raingauge stations, equipped with ordinary raingauges, cover the Bina river watershed area, namely Begamganj, Gairatganj, Rahatgarh and Silvani stations. The study area falls under the Vindhyan region, important rocks found in this area are sand stone, Quartizitic sand stone, lime stone and Deccan traps, called basalt. The topography of the area is rolling to undulating. The valley land is moderately to poorly drain. The average normal annual rainfall of the area is 1204.5 mm and about 90% of the annual rainfall takes place during the Southwest monsoon period i.e. June to October. The area around Bina river basin is mostly fertile black cotton soil and some area is under red soil. The main crops grown in Kharif season are soyabeen, urad and paddy and main crops grown in Rabi season are wheat and red gram. METHODOLOGY Thornthwaite and Mather s model Analysis of water availability in the Bina river watershed is based on the Thornthwaite methodology. This method mainly calculates potential evapotranspiration (PET). According to the PET results and using Thornthwaite and Mather s model, water storage, storage changes, actual evapotranspiration, surplus and deficit can be calculated. The inputs are temperature, precipitation, day length hours as a function of latitude and available water capacity of the soil. Thornthwaite (1948) correlated mean monthly temperature with evapotranspiration as determined from water balance for valleys where sufficient moisture water was available to maintain active transpiration. In order to clarify the existing method, the Thornthwaite s equation is discussed, L N 10 T PET 16 a (1) 12 30 I Where PET is the estimated potential evapotranspiration (mm/month) T a is the average daily temperature (degrees Celsius; if this is negative, use 0) of the month being calculated N is the number of days in the month being calculated L is the average day length (hours) of the month being calculated ( 6.75 10 7 ) I 3 (7.71 10 5 ) I 2 (1.792 10 2 ) I 0.49239 (2) 27

1.514 12 T I ai (3) i 1 5 As per the Thornthwaite and Mather s model next step is to calculate (P PET), which is an estimation of the quantitative water excess (+) or deficit ( ) with P as precipitation. Then the accumulated values of (P PET), i.e. the accumulated potential water loss (APWL) for each month, are calculated. This will be zero for months having positive (P PET) and starting with the first month having a negative value after the monsoon. Then the actual storage of soil moisture (STOR) for each month is calculated as follows: ( APWL / AWC ) STOR AWC e (4) where, AWC is the moisture storage capacity (i.e. the available water capacity) of the soil. This is calculated based upon the land use, soil texture and rooting depth as suggested by Thornthwaite & Mather (1955, 1957). Changes of actual storage ( SM) for all the months are calculated as: SM month STOR month STOR (5) previous month A negative value of SM implies subtraction of water from the storage to be used for evapotranspiration, whereas a positive value of SM implies infiltration of water into the soil and its addition to the soil moisture storage. The actual evapotranspiration (AET) is computed for all the months, as given in the following equations; AET = ΔSM + P ; ΔSM < 0 (6) AET = PET ; ΔSM > 0 (7) The water deficit (DEF) for crop evapotranspiration in each month is calculated for the months having negative (P PET) as follows: DEF PET AET (8) The amount of excess water that cannot be stored is denoted as moisture surplus (SUR). When storage reaches its capacity, SUR is calculated using equation given below: SUR = P PET (9) When the soil storage is not at its capacity, no surplus exists. In a month in which the soil moisture storage capacity is just satisfied, SUR is obtained using equation given below: SUR = P (AET + SM) (10) Where, ΔSM is the change in actual soil moisture storage. The runoff is calculated using the formulae given below: R. Omonth month previousmonth (0.5 SUR ) (0.5 SUR ) (11) Where, R.O is runoff for any month. The annual amount of actual evapotranspiration and runoff from the watershed is calculated considering the area under different land use and the respective values from the monthly water balance table. Thus, the monthly actual evapotranspiration and runoff from the watershed are areaweighted values. DATA USED The data required for the study are mainly climatic, hydrologic and the basin characteristics, viz. watershed boundary, soil texture, landuse class, temperature, day length, rainfall, runoff, etc. The precipitation and discharge data of the Bina river watershed was available for very limited period, 2007 to 2010. Extent of the Study Area The Bina river watershed up to Rahatgarh is covered by 1:50,000 scale Survey of India (SOI) toposheet no. 55-I / 2, 3, 6, 7 and 11 on 1:50,000 scale, which have been used for digitizing the catchment boundary, drainage line and location of rain gauge stations, using ILWIS 3.0 software. The Indian Remote Sensing Satellite Resourcesat-1, LISS-III sensor and Path 98 -Row 56 covering the entire catchment area has been used for preparation of first level Landuse/land cover map. Meteorological Data As the meteorological data such as temperature and sunshine hours do not change much in a time span of few years, so the average values of meteorological data, for the period of 1 st January, 1972 to 31 st December, 2003 of Sagar observatory collected from the Indian Meteorological Department (IMD), Pune have been used in Thornthwaite equation for computation of monthly values of PET. Rainfall and Discharge Data The daily rainfall data of four rain gauge stations, namely Begamganj, Gairatganj, Rahatgarh and Silvani falling in and around the study area for the period of 1 st January, 2007 to 31 st December, 2010 were available with State Water Data Centre, Water Resources Department (Govt. of M.P.), Bhopal and Superintendent of Land Record, Sagar. The daily gauge and discharge data observed at Rahatgarh (G/D) site on Bina river for the same period, i.e. 1 st January, 2007 to 31 st December, 2010 were collected from State Water Data Centre, Water Resources Department (Govt. of M.P.), Bhopal. RESULTS AND DISCUSSION A number of data such as climatic data, soil and land use/land cover data from various sources encompassing various vistas have been integrated in GIS environment to generate output for water resource planning projects to overcome the problems of water scarcity, over exploitation of available water resources and also to make use of run-off by suitable storage measures to reduce dependence on ground water resources in Bina watershed. MS Excel worksheet was used for all mathematical computations using the equations discussed in methodology section to obtain PET, AET, APWL, surplus, deficit, soil moisture etc. The ILWIS GIS platform has been used for preparation of thematic maps and map calculations using the watershed area, soil and landuse maps. Various outputs were generated in both tabular and map forms. Rainfall The observed rainfall have been accumulated for monthly values, since the water balance computation has been done on monthly basis. The Thiessen Polygon map was created from 28

the point map in raster format in GIS platform and then the weights of each influencing stations in the study area were estimated based on their areal extent. The weighted average monthly precipitations over the watershed have been computed. Soil The soil map published by the National Atlas & Thematic Mapping Organization (NATMO), Department of Science & Technology, Government of India, Kolkata was digitized and stored in vector format (Fig. 2). Various categories of soils are found in the study area, namely fine loam, clayey loam, loamy clay and fine clay. The information on soil classification was taken from the reports Soils of Madhya Pradesh for Optimising Land Use and Soil Series of India published by National Bureau of Soil Survey and Land Use Planning (ICAR), Nagpur. prepared by applying Maximum Likelihood Classifier (MLC) digital classification of multi-date LISS-III satellite data acquired from the National Remote Sensing Centre (NRSC), Hyderabad. Using Maximum Likelihood Classifier (MLC) nine land use classes were identified in Bina watershed namely dense forest, open forest, agriculture land 1 (wheat), agriculture land 2 (mustard/gram), current fallow, barren, settlements, water body and land with shrubs, which is given in Fig. 3. Figure 2: Soil Map of Bina River Watershed Figure 3: Landuse Map Bine River Watershed Water Balance Calculation The basic assumption made in the Thornthwaite & Mather method is that the surplus and deficit moisture in the soils in root zone control the actual evapotranspiration. If the available moisture > PET, than the AET equals the PET, otherwise the AET is limited to the available moisture, i.e. AET < PET. The Table 1: Computation of accumulated potential water loss during 2010 (values in mm) Month Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Total P 0.78 0.60 0.00 0.00 0.00 103.90 220.67 230.58 110.91 78.14 38.30 0.00 783.90 PET 33.96 51.20 95.05 176.24 321.63 169.73 72.00 54.12 81.22 94.05 51.62 28.50 1229.31 P-PET -33.17-50.60-95.05-176.24-321.63-65.82 148.68 176.46 29.69-15.91-13.32-28.50-445.41 APWL -90.90-141.50-236.55-412.79-734.42-800.25 0.00 0.00 0.00-15.91-29.23-57.73-2519.27 Land Use The ILWIS GIS software has the capability of Digital Image Processing (DIP) capabilities also. The landuse map was available moisture (rainfall), the PET and AET have been computed by using the equations suggested by Thornthwaite and Mather method. The surplus moisture is responsible for 29

runoff in the river basin, which flows either as direct runoff or base flow. In the present study, these computations have been made in MS Excel sheets and compiled to get the monthly volumetric runoff in MCM. For computing the climatic water balance using the TM model, monthly potential evapo-transpiration was calculated using the eq.1. Next, P PET, which is an estimation of the quantitative water excess (+) or deficit ( ) with P as precipitation was calculated. Then the accumulated values of (P PET), i.e. the accumulated potential water loss (APWL) for each month, J. Indian Water Resour. Soc., Vol 35, No. 3, July 2015 Table 2: Computation of water holding capacity of the root zone and available water capacities (AWC) for different soil textures and land uses Landuse * Soil class Area (Sq.km) FC (%Vol.) were calculated. The results given in Table 1 show that the annual rainfall in Bina watershed during the year 2010 was 783.90 mm, PET was 1229.31 mm and maximum value for APWL was found to be 800.25 mm in June. Available Water Capacity (AWC) AWC is the moisture storage capacity of the soil. For water balance calculations, the total available water holding capacity in a soil profile should be known. A large water-holding capacity implies a small annual runoff and a large annual evapotranspiration relative to a small water-holding capacity PWP (%Vol.) AWC (%Vol.) Rooting depth (m) AWC of root zone (mm) Dense forest * Fine Loam 2.09 31 11 20 1.5 300 Dense forest * Clayey Loam 2.38 36 22 14 1.5 210 Dense forest * Loamy Clay 6.72 27 17 10 1.5 150 Dense forest * Fine Clay 23.83 42 30 12 1.5 180 Open forest * Fine Loam 23.26 31 11 20 1 200 Open forest * Clayey Loam 14.21 36 22 14 1 140 Open forest * Loamy Clay 33.41 27 17 10 1 100 Open forest * Fine Clay 224.81 42 30 12 1 120 Agriculture1 * Fine Loam 3.57 31 11 20 0.62 124 Agriculture1 * Clayey Loam 2.91 36 22 14 0.62 86.8 Agriculture1 * Loamy Clay 0.96 27 17 10 0.62 62 Agriculture1 * Fine Clay 18.58 42 30 12 0.62 74.4 Agriculture2 * Fine Loam 22.52 31 11 20 0.38 76 Agriculture2 * Clayey Loam 10.26 36 22 14 0.38 53.2 Agriculture2 * Loamy Clay 6.18 27 17 10 0.38 38 Agriculture2 * Fine Clay 120.41 42 30 12 0.38 45.6 Current fallow * Fine Loam 23.74 31 11 20 0.62 124 Current fallow * Clayey Loam 9.24 36 22 14 0.62 86.8 Current fallow * Loamy Clay 13.72 27 17 10 0.62 62 Current fallow * Fine Clay 165.12 42 30 12 0.62 74.4 Shrub land * Fine Loam 18.00 31 11 20 0.75 150 Shrub land * Clayey Loam 13.28 36 22 14 0.75 105 Shrub land * Loamy Clay 26.99 27 17 10 0.75 75 Shrub land * Fine Clay 259.10 42 30 12 0.75 90 barren land * Fine Loam 3.18 31 11 20 1 200 barren land * Clayey Loam 1.60 36 22 14 1 140 barren land * Loamy Clay 5.55 27 17 10 1 100 barren land * Fine Clay 44.90 42 30 12 1 120 Settlement * Fine Loam 4.51 31 11 20 0.5 100 Settlement * Clayey Loam 1.68 36 22 14 0.5 70 Settlement * Loamy Clay 2.03 27 17 10 0.5 50 Settlement * Fine Clay 25.54 42 30 12 0.5 60 30

Values in mm J. Indian Water Resour. Soc., Vol 35, No. 3, July 2015 under the same climatic conditions. This has been calculated based upon the land use, soil texture and rooting depth as suggested by Thornthwaite & Mather (1955, 1957). The computations have been summarised in Table 2, which gives the respective areas covered by different soils in each land use class and rooting depth used to calculate the available water capacity (AWC) of root zone, based on the field capacity (FC) and permanent wilting point (PWP) of the respective soil class. Table 3: Computation of P, PET, AET and Runoff for year 2010 (values in mm) Month Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Total P 0.78 0.60 0.00 0.00 0.00 103.90 220.67 230.58 110.91 78.14 38.30 0.00 783.90 PET 33.96 51.20 95.05 176.24 321.63 169.73 72.00 54.12 81.22 94.05 51.62 28.50 1229.31 AET 15.31 15.06 13.74 8.17 2.52 104.02 72.00 54.12 81.22 92.65 48.56 17.40 524.75 R.O 0.00 0.00 0.00 0.00 0.00 0.00 27.63 114.72 101.93 14.85 0.00 0.00 259.19 350 300 250 200 Deficit Soil Moisture Recharge Surplus 150 100 Soil Moisture Utilization Soil Moisture Utilization 50 0 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Months P PET AET Runoff Figure 4: Water balance status of the Bina river watershed (2010) 160.00 140.00 120.00 100.00 80.00 60.00 40.00 20.00 0.00-20.00 R.O (obs.) MCM R.O(cal.) MCM Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Figure 5: Comparison of observed and modeled runoff (2010) 31

Year P (mm) Table 4: Summary of calculations for the year 2007-2010 AET (mm) Obs. runoff (mm) Cal. runoff (mm) DEF (mm) SUR (mm) Nash - Sutcliffe (E %) Relative vol. error (RVE %) 2007 781.96 402.20 422.28 379.76 827.11 379.76 0.77-10.5 2008 935.06 516.79 378.36 418.27 712.53 418.27 0.76 10.5 2009 1182.51 524.28 638.58 658.27 705.03 658.23 0.78 03.1 2010 783.94 524.75 279.45 259.19 704.61 259.19 0.97-07.2 Surface Runoff The monthly water balance consisting the variation of precipitation (P), potential evapo-transpiration (PET), actual evapotranspiration (AET) and runoff in the Bina river watershed has been calculated using equation 1 to equation 11. The monthly variation of these parameters in the watershed has been presented in Table 3. It gives a great deal of information regarding the water balance of the Bina watershed and also total runoff. Besides showing the seasonal pattern of precipitation, actual evapotranspiration (AET), potential evapotranspiration (PET) and runoff, Figure 4 indicates the periods of moisture deficit and soil moisture recharge. The moisture deficit indicates that plants will be under some stress during that period, indicating the need for irrigation. Finally a comparison between observed and calculated runoff is shown in Figure 5 to show the accuracy of the calculated run-off and the peak discharge. CONCLUSION A mathematical model which is based on Thornthwaite and Mather s water balance methodology is prepared for assessing water balance in Bina watershed. Initially 1 year data (January 2010 to December 2010) is used for developing this model. Then the accuracy of model has been checked for 3 more individual years viz., 2007, 2008 and 2009. Table 4 shows the summary of calculations for all the four years. The study reveals that the area is relatively dry in the months of January June. Further, it can be seen that there is a water deficit again after monsoon in the months of October December. Soil moisture recharge takes place between July and August. From July to September is the period of water surplus. The results of efficiency test carried out for the model shows that the Nash-Sutcliffe Coefficient (E) varies between 0.76 to 0.97 and the Relative Volume Error (RVE) ranges from ±10 % for all the four years for which surface runoff were computed. The table below shows the computed values of efficiency and relative volume error. ACKNOWLEDGEMENTS The authors owe many thanks to the Director, National Institute of Hydrology, Roorkee and the Coordinator, NIH Regional Centre, Bhopal and the authors also owe their gratefulness to the Director, MANIT and the HOD, Civil Engineering Department, MANIT, Bhopal for providing institutional facilities to complete this work. REFERENCES 1. Jain, Sharad K., 2012. India s water balance and evapotranspiration, Current Science, Vol. 102, NO. 7, pp. 964-976. 2. Jasrotia, A.S., Majhi, Abinash and Singh, Sunil, 2009. Water Balance Approach for Rainwater Harvesting using Remote Sensing and GIS Techniques, Jammu Himalaya, India, Water Resource Manage 23, pp. 3035 3055. 3. Jenifa Latha, C., Saravanan, S. and Palanichamy, K., 2010. A semi-distributed water balance model for Amaravathi river basin using remote sensing and GIS, Int. Journal of Geomatics and Geosciences, Volume 1, No 2, pp. 252-263. 4. Karsili Cansu, 2013. Calculation of past and present water availability in the Mediterranean Region and future estimates according to the Thornthwaite waterbalance Model, Department of Physical Geography and Ecosystems Science, Lund University. 5. Killingtveit Anund, 2004. Water balance studies in two catchments on Spitsbergen, Svalbard and Northern Research Basins (Proceedings of a workshop held at Victoria, Canada, March 2004). IAHS Publ. 290, pp. 120-128. 6. Kumar, S.R., 2001. Water balance study of Krishnai River Basin according to Thornthwaite s concept of potential evapotranspiration, National Institute of Hydrology, Roorkee. 7. Nayak, T.R., 2013. A Status Report on Pilot Basin Studies: IWRM in Bina River Basin in Bundelkhand Region in Madhya Pradesh, National Institute of Hydrology, Roorkee. 8. Pimenta M.T., 1999. Water Balances Using GIS, Phys. Chem. Earth (B), vol. 25, No. 7-8, pp. 695-698. 9. Roy, Sagarika and Ophori, Duke, 2012. Assessment of water balance of the semi-arid region in southern San Joaquin Valley California using Thornthwaite and Mather.s model, Journal of Environmental Hydrology, Volume 20, Paper 15. 10. Singh R.K, Hari Prasad V, Bhatt C.M., 2004. Remote sensing and GIS approach for assessment of the water balance of a watershed. Hydrological Sciences 32

Journal des Sciences Hydrologiques, vol. 49(1), pp. 131-141. 11. Subrahmanyam V.P., 1983. Water balance approach to the study of aridity and droughts with special reference to India, New Approaches in Water Balance Computations (Proceedings of the Hamburg Workshop, August 1983). IAHS Publ. no. 148, pp. 125-135. 12. Thornthwaite C.W. 1948. An approach toward a rational classification of climate. Geographical Review 38: 55 94 13. Thornthwaite C.W. & Mather, R. J., 1955. The water balance: Publications in Climatology 8, 1 86. DIT, Laboratory of Climatology, Centeron, New Jersey, USA. 14. Thornthwaite, C.W., and J.R. Mather., 1957. Instructions and tables for computing potential evapotranspiration and the waterbalance: Centeron, N.J., Laboratory of Climatology, Publication in Climatology, Vol. 10, no. 3, pp. 185-311. 15. Victoria, Santiago, Ballester, Pereira, Richey (2006). Water balance for the Ji-Parana River Basin, Western Amazon, using a simple method through Geographical Information Systems and Remote Sensing. 33