Sudheer, K P1,2,*, S. Murty Bhallamudi1,3, Balaji Narasimhan1,3, Jobin Thomas1, Bindhu, V M1, Vamsikrishna Vema1,4, Cicily Kurian1,
1Department of civil Engineering, Indian Institute of Technology Madras, Chennai – 600036,
2Department of Agricultural and Biological Engineering, Purdue University, West Lafayette, IN, USA.
3Indo German Centre for Sustainability, Indian Institute of Technology Madras, Chennai – 600036
*Corresponding Author: email@example.com
The authors of Sudheer et al. (2019) – hereinafter referred to as ‘authors’ – appreciate Mr. J. Harsha (hereinafter referred to as ‘commenter’) for his judgmental assessment (in his blog appeared on SANDRP website- “https://sandrp.in” on August 25, 2020, see: https://sandrp.in/2020/08/25/role-of-dams-in-kerala-floods-distortion-of-science/) of the authors’ work “Role of dams on the floods of August 2018 in Periyar River Basin, Kerala” (published in the Current Science in 2019: [DOI: 10.18520/cs/v116/i5/780-794]). As mentioned in Sudheer et al. (2019), the primary objective of the article was to examine whether the early release of the water stored in the reservoirs would have attenuated the flood peaks, and if so, what would have been the extent of the attenuation, in the context of debates and discussions in the social, political, as well as scientific domains based on non-sequitur speculations. Accordingly, the authors designed a scientific exercise using a widely used hydrological model (HEC-HMS) to understand the role of the dams in the Periyar River Basin (PRB) in the 2018 flooding situation.
The commenter raised a few concerns regarding the data, methodology and assumptions that the authors used in Sudheer et al. (2019). The authors discuss and illustrate below that there is no flaw in the methodology, modelling, results and analysis presented in Sudheer et el. (2019), and reiterate that the inferences outlined in the conclusion section of the article by Sudheer et al. (2019) were clearly based on the results presented and discussed in different sections of the article. These inferences were arrived at based on scientific evidences through simulation, and there is no distortion of science. The authors clearly demonstrate that the commenter’s concerns amount to ‘judgemental assessment’. They are shallow, misguided by the intuitions mistaken for science, and miss the point we made by several miles.
- Is SCS-CN method valid for Periyar River Basin (PRB)?
The Soil Conservation Service Curve Number (SCS-CN) method was developed in 1954, and is documented in Section 4 of the National Engineering Handbook (NEH-4) of the U.S. Department of Agriculture in 1956 (Mishra and Singh, 2003). The document has since been revised in 1964, 1965, 1971, 1972, 1985, and 1993. Note that the SCS-CN method is an outcome of exhaustive field investigations carried out during the late 1930s and early 1940s, in several experimental catchments distributed across different hydro-climatological conditions, and the works of several early investigators, including Mockus (1949), Sherman (1949), Andrews (1954), and Ogrosky (1956). The SCS-CN method is simple, easy to understand and apply, stable, and useful for ungauged watersheds with only minimal data. The primary reason for its wide applicability and acceptability across the globe in different geological, geomorphological, climatic and hydrological conditions lies in the fact that it accounts for most runoff producing watershed characteristics: soil type, land use/treatment, surface condition, and antecedent moisture condition (Mishra and Singh, 2003). The SCS-CN method is based on the water balance concept and two fundamental hypotheses. The first hypothesis equates the ratio of the actual amount of direct surface runoff (Q) to the total rainfall (P) (or maximum potential surface runoff) to the ratio of the amount of actual infiltration (F) to the amount of the potential maximum retention (S). The second hypothesis relates the initial abstraction (Ia) to the potential maximum retention. The parameter, λ, relates Ia and S. In general, λ varies between 0 and 0.30. The value of λ is sensitive primarily for low flows and needs careful consideration if the interest is to simulate infiltration and ground water recharge for a long-term simulation run. However, in the case of Sudheer et al (2019), where simulation of high flows are performed with wet antecedent conditions only for the 2018 event, and in the premise that the curve number is automatically adjusted for wet conditions within HEC-HMS, an average value of λ = 0.2 is valid. Furthermore, the sensitivity of λ, as explicitly demonstrated (in Page No 120; Mishra and Singh, 2003), is that the smaller the value of CN, the larger the effect of variation of λ on C (i.e., the runoff factor, Q/P). On the other hand, when CN and P are large, the corresponding C-value is also large, but insensitive to the variation of λ. Therefore, the assumption of λ = 0.2 in PRB during the AMC-II/III conditions is valid.
2. Whether land use/land cover for the year 2010 is valid for August 2018?
It is true that the authors used the land use/land cover data corresponding to the year 2010, obtained from National Remote Sensing Centre (NRSC), as was mentioned in the article (Sudheer et al., 2019; Page 783-784). An investigation into the land use changes indicated that the vegetative cover has only decreased by 0.50% in 6 years in PRB (4079.07 km2 in 2010 as compared to 4058.78 km2 in 2016, according to NRSC data corresponding to respective years). It is also noted that there is an increase in the built-up area of the Periyar River Basin (16.51%; 56.93 km2 as compared to 66.33 km2 in 2016), which was visible mostly in the downstream of Neeleswaram. To the best of the knowledge of the authors, there were no significant land use changes in the upstream areas (where most part of the runoff is generated) of PRB after 2010, and therefore, using the land use/land cover data pertaining to the year 2010 for the study does not induce large uncertainty in the simulations, especially during heavy flooding scenario. The authors wish to mention that the areal extent of any land use class would depend on the resolution of satellite image, the designated number of classes and the classification algorithm used for classifying the satellite image, and therefore a direct comparison of the classification performed by the commenter and by the NRSC may not be appropriate.
3. Is PRB dendritic?
As the commenter mentioned, the basin supports diverse drainage patterns at various spatial scales. The major drainage patterns of the basin include dendritic, parallel, trellis and rectangular, where dendritic pattern was observed in gentle slope areas irrespective of the physiography. The parallel pattern is obvious in high slopes and trellis and rectangular patterns infer the structural controls. However, we followed the observations made by Soman (2002, Page 7-9) in “Geology of Kerala” that the dominant drainage pattern of the region is dendritic.
4. Quality of discharge data for validation of HEC-HMS simulations
Central Water Commission (CWC) is the premier Technical Organization of India in the field of Water Resources, and functions as an attached office of the Ministry of Jal Shakti, Department of Water Resources, River Development and Ganga Rejuvenation, Government of India (http://www.cwc.gov.in). One of the several functional domains of activities of CWC is River Management, which is responsible for collection, compilation, storage and retrieval of hydrological and hydro-meteorological data including water quality monitoring (http://www.cwc.gov.in). Therefore, any researcher could rely on the data provided on public domain by the CWC for any study related to water resources development and management in the country. It is to be noted that the authors’ used discharge data at Neeleswaram during August 2018 for validation, which was published by the CWC (CWC, 2018), and is clearly mentioned in the article. The commenter has probably wrongly discerned that the study used the entire discharge data during 1971-2017. As long as the CWC does not mention any remarks about the unreliability of the data while publishing on public domain, one need not suspect the authenticity of the data. However, the commenter’s remarks that the data provided by the CWC is unreliable, is a concern that is to be addressed to the CWC and the Government directly, as it is purely a policy issue. It is more ironical that the commenter, who himself is a Director of CWC and in a leadership position to make policy of data acquisition, quality control and dissemination, has questioned the very reliability of their own data.
5. Lateritic soils classified as HSG-D soil is an error
Sudheer et al. (2019) mentioned that four types of soil are seen in PRB, viz. forest soil, hill soil, laterite soil and alluvial soil. The upstream areas of the basin are dominantly covered by forest and hill soils, whereas the downstream parts are covered by alluvial soil. The laterite soils interspersed with alluvial soils are dominated in the midlands and parts of downstream areas of the basin (Page No. 784). This is the general distribution of the different soil types. However, the attribution of the hydrological soil group (HSG) was not done based on the general soil type, instead, we used the soil texture given by the soil data of NBSS & LUP. The HSG “D” was assigned to clayey soils only, whereas the clay and loam rich soils are assigned as C. The A and B groups are attributed to gravel and sand dominated textures. Hence, it seems that the commenter raised this concern based on his assumptions or judgment.
6. Where are the reservoir rule curves?
Sudheer et al. (2019) followed the standard operating policy for conservation storages in the simulations for all the dams. The issue about the rule curve or the lack of it have been clearly flagged in the manuscript in pages 790, 792 and 793, which the commentator seems to have conveniently ignored. Having, a rule curve alone will not solve the flooding issue. It needs to be used in conjunction with a good inflow forecasting system coupled with telemetry system updating the actual measurements at frequent interval. The need for such a forecast system has been clearly brought out in page 792 in Sudheer et al. (2019).
The commenter, identified the inferences made by Sudheer et al. (2019), as follows, and argued that these inferences are not appropriate in the context of the assumptions and the poor quality data employed. The discussions presented against each of the points raised by the commenter, as elucidated above, clearly illustrates that there is no flaw in the methodology, modelling, analysis and results presented in Sudheer et el. (2019). Despite, the following information makes the inferences from the study much more clear.
- The reported flows at downstream locations will be insensitive to the storage status of the reservoirs, if the reservoir is filled to less than 50% of its capacity.
Sudheer et al. (2019) have explicitly mentioned in the conclusions regarding the flow reductions in the downstream (i.e., Neeleswaram) that are possible against different storage status in the upstream reservoirs (Page No. 793). It was mentioned that the reservoir storage of 75% of the total storage by the end of July 2018 can reduce 16% of the actual peak flow, while a controlled storage of 50% or 25% suggested the peak attenuation by 21%. It may be noted that the gross storage capacity of the Idukki reservoir is 1996 MCM at FRL (2,403 ft), whereas the 50% and 25% storages correspond to ~2,335 ft and ~2,270 ft, respectively. However, the crest level of spillway of Idukki dam is 2,373 ft. Similarly, the gross storage of Idamalayar reservoir is 1090 MCM at FRL (169 m), whereas the 50% and 25% storages indicate the water levels at ~149 m and ~135 m, respectively. However, the crest level of spillway is at 161 m. Hence, there won’t be any spillway discharge for 50% and 25% reservoir storage levels. Since the difference in the storage volume between the 50% and 25% storage levels is relatively less, the heavy reservoir inflows during the EREs will make it up within shorter periods (i.e., within a day or two), which could be the primary reason for the insensitivity in the flood flow reduction for storages lesser than 50%. It is worth mentioning here that the PRB received different spells of heavy rainfall prior (during 6-11, August 2018) to the devastating rainfall event (15-17, August 2018) in discussion.
2. The bank full discharge at L1 and L2 approximately being 3,400 m3/s and 4,200 m3/s, it is apparent that the reservoir operation could not have helped in avoiding the flood situation.
It is true that we have mentioned: “with the bank full discharge at Neeleswaram approximately being 4,200 m3/s, it is apparent that the reservoir operation could not have helped in avoiding the flood situation”. The commenter may please construe the premise on which such a statement is made; Sudheer et al. (2019) have examined several scenarios including a scenario where all the reservoirs were at 25% of its capacity by the end of July 2018 (Table 2, Run No. 8, Page 786). This run specifically has a provision to accommodate the flood storage which helps in attenuating the peak flows. The results of the run indicated that the peak flow at Neeleswaram was 7,844 m3/s, which is 86% higher than the bankfull discharge at Neeleswaram. It is evident from Fig. 12 that the simulated/observed peak flows at Neeleswaram were much higher as compared to the bankfull discharge at the location. Therefore, it is obvious that any reservoir operations could not have avoided the flooding situation. Note that the bank full discharges at L1 and L2 were estimated using the principles of hydrologic and hydraulic computations.
3. A quantitative comparison indicates that the hydrologic response of the major tributaries of the Periyar River was also significant in terms of magnitude of flows, compared to the reservoir releases.
It was inferred from the simulations that the contributions from the major tributaries were more than the releases from the reservoir. Note that the distributed computing nature of the HEC-HMS facilitates drawing the discharges from every sub-basin outlet. In Page 788, it is clearly mentioned that the flood discharges from one of the nearly uncontrolled tributaries, Perinjankutty, itself had a discharge of 3,500 m3/s as compared to 1,860 m3/s from Idukki reservoir. Please note that in addition to Perinjankutty, there were other sub-basins, which also contributed higher flood flows. The flood discharges of all the major sub-basins were estimated from the simulations. The spatial distribution of rainfall is a significant factor in runoff generation, as is clearly stated in the manuscript and depicted in Figure 4. Hence, huge amount of inflow primarily happened from the uncontrolled tributaries, rendering any flood cushion created in the dams with early release was ineffective for flood control. Uncertainty in the spatial distribution of rainfall can be addressed only by installing more automatic weather stations and doppler radar network in the catchment, and in conjunction with the inflow forecasting system and appropriate rule curve.
4. A major share of the peak flow at L2 is contributed by the catchment area of the major tributaries, such as Perinjankutty, Muthirapuzha and Idalamalayar.
Please see the response to point # 3 above.
5. Compared to virgin simulations, which resulted in peak flow magnitudes 8,224 m3/s and 11,990 m3/s at L1 and L2 respectively, the reservoirs were effective in reducing the peak flows to the scale of 2500 m3/s.
Compared to the virgin simulations, implying no dams present in the PRB, (peak flow magnitude at L1=8,224 m3/s, L2=11,990 m3/s), the reservoirs were effective in reducing the peak flows (L1 = 5,523 m3/s, L2 = 9,965 m3/s), which showed a reduction on an average of 2,500 m3/s just because of the mere presence of dams as compared to virgin basin.
All these inferences, outlined in the conclusion section of the article by Sudheer et al. (2019), were clearly based on the results presented and discussed in earlier sections of the article, and not derived or inferred without any scientific evidence, and there is no distortion of science. The author’s wish to make additional information here about the actual flood situation in 2019 in the PRB, which the commenter might have come across. Note that the water level in the Periyar River had risen to alarming levels by August 9, 2019, while there was no release from any of the major dams. The water level at Idukki dam was 30% of FRL, and at 40% of FRL in the case of Idamalayar dam. Note that the River was overflowing in August 2019, leaving several low-lying areas inundated, including the Cochin International Airport, which was shut down for operations. This situation conforms the inferences of Sudheer et al. (2019) that even without release from the major dams, due to the spatial variability of the rainfall, the hydrologic response of the catchment area of Periyar River results in swelling of the River to precarious levels. It clearly shows that the commenter’s concerns amount to ‘judgemental assessment’. They are shallow, misguided by the intuitions mistaken for science, and miss the point we made by several miles.
Andrews, R.G. (1954). The use of relative infiltration indices for computing runoff (unpublished). Soil Conservation Service, Fort Worth, Texas, 6p.
CWC (2018). Kerala Floods of August 2018. Central Water Commission, New Delhi.
Mishra, S.K., Singh, V.P. (2003). Soil conservation service curve number (SCS-CN) methodology. Water Science and Technology Library, Volume 42. Springer, Dordrecht.
Mockus, V. (1949). Estimation of total (peak rates of) surface runoff for individual storms. Exhibit A of Appendix B, Interim Survey Report Grand (Neosho) River Watershed, USDA.
Ogrosky, H.O. (1956). Service objectives in the field of hydrology. Unpublished, Soil Conservation Service, Lincoln, Nebraska, 5pp.
Sherman, L.K. (1949). The unit hydrograph method. In: O.E. Meinzer (ed.), Physics of the Earth, Dover Publications, New York, pp. 514-525.
Sudheer, K.P., Bhallamudi, S.M., Narasimhan, B., Thomas, J., Bindhu, V.M., Vema, V., & Kurian, C. (2019). Role of dams on the floods of August 2018 in Periyar River Basin, Kerala. Current Science, 116(5), 780-794. doi:10.18520/cs/v116/i5/780-794.