(1985) Medina River Fish Kill

--- Page 1 --- MEDINA RIVER FISH KILL BY Roy W. Spears Texas Parks and Wildlife Department Resource Protection Division Pollution Surveillance Branch May 1, 1986 --- Page 2 --- INTRODUCTION A Southern Pacific train consisting of 112 cars derailed at 10:40 p.m. on September 14, 1985. The wreck occurred at the Medina River bridge near the town of Macdona in southwestern Bexar County (Figure 1.) The train was transporting 112 tank cars including 21 cars that were carrying approximately 286,000 gallons of 95 percent sulfuric acid. When the tank cars ruptured, an estimated 264,000 gallons of sulfuric acid were released. Approximately 42,000 gallons flowed from the spill site down Nelson Road and entered the Medina River in the vicinity of a bridge known as the low water crossing. Approximately 50,000 gallons of acid were retained by emergency berms, dikes and natural depressions and were collected and transported off site. Approximately 22,000 gallons of acid remained in the tank cars and were removed for disposal or reclamation. The balance of material either soaked into the soil or flowed into an abandoned gravel quarry immediately adjacent to the spill site. Sulfuric acid disassociates almost completely in water to release sulfate ions and hydrogen ions. An increase in hydrogen ion concentration would lower the pH. A pH range of 6.5 to 9.0 has been established to provide adequate protection for freshwater fish and invertebrate fish food organisms. Outside of this range, fish suffer adverse physiological effects increasing in severity as the degree of deviation from this range increases until lethal levels are reached (EPA, 1976). According to the American Fisheries Society (1979), a pH range of 6.5 to 8.5 with no change greater than 0.5 units outside estimated natural seasonal maxima and minima provides nearly a maximum level of protection for freshwater aquatic life. The toxicity of sulfuric acid to aquatic life is a function of the resulting pH (Doudoroff and Katz, 1950). Ellis (1943) confirmed this observation by stating that sulfuric acid must be diluted to give a pH of 4.5 or higher in order for fish to survive, and to pH 5.5 for other aquatic organisms to thrive and provide food for fish. Sunfish, bass and carp exhibit gill irritation and/or death below pH 4.0 (Lewis 1956). When sulfuric acid in streams depresses the pH below 5.0, atypical aquatic ecosystems develop (Harrison 1958). Also, species diversity declines and only a few resistant species remain which dominate the system (Boesch 1977). Due to the large quantity of sulfuric acid that reached the river, Texas Parks and Wildlife Department (TPWD) made an investigation of the spill to assess damage to aquatic life in the Medina River. Law Enforcement personnel were on the scene within two hours after the spill occurred. Restricted access to the site and concern for public safety delayed fish kill investigation activities until the evening of September 15, 1985. MATERIALS AND METHODS Five 150 foot segment counts were made on September 16 by Law Enforcement and Resource Protection personnel between the spill site at FM 1604 and TH-35, a distance of 7.2 miles (Figure 1). Sampling sites were selected 0-142A-05/01 /86 --- Page 3 --- on the basis of their accessibility. Due to the harmful acidic condition of the river, only shoreline counts were made. All of the fish counted were identified and measured by inch class. Seven 150 foot segment counts were made on September 18 by Inland Fisheries personnel between the spill site and Applewhite Road, a distance of 17 miles. These segments were also selected according to accessibility. Two boats were launched at IH-35 and motored upstream and downstream to select count segments based on bottom type and river flow; j.e., deep holes, riffle areas, fast flowing water and low flowing water. Segment lengths were determined by anchoring a 150-foot rope at the lead end with a float attached at the trail end. All fish along the shoreline and in the river within this 150 foot distance were identified, counted and measured. On September 20, ten 150 foot segments were counted by Inland Fisheries and Resource Protection personnel from Applewhite Road downstream to FM 1937, a distance of 12 miles. Two boats were launched at the U.S. 281 bridge. One crew made three segment counts. The first sampling segment was established by the sampling crew by motoring upstream for 25 minutes. The second and third segments were selected by floating downstream at four minute intervals. The second crew made seven segment counts, spacing the distance between segments at four minute intervals. Each segment distance was measured with a 150 foot rope anchored at the lead end and a float attached at the trail end. All species were counted, measured and identified by inch class. The number of fish in each inch class for each species were summed and multiplied by an expansion factor to obtain an estimate of the total number of dead fish in the river. Expansion factors were calculated according to the following general equation: EF = total stream miles affected summed lengths of segments counted Expansion factors were calculated for each day counts were made. These expansion factors are: (5,280 ft/mile) (7.2 miles) September 16, 1985 EF = S (150 ft.) (Spill site to IH-35) EF = 50.7 5,280 ft/mile) (17 miles September 18, 1985 EF = 7 (150 ft.) (Spill site to Applewhite Rd.) EF = 85.5 5,280 ft/mile) (12 miles September 20, 1985 EF = 10 (150 ft. (Applewhite Rd. to FM 1937) EF = 42.2 0-142A-05/09/86 --- Page 4 --- TPWD personnel made pH measurements with a portable Hach meter. Qualitative observations of water color and turbidity were made at all river crossings from the spill site to its confluence with the San Antonio River. Additional pH data were obtained from the Texas Water Commission and San Antonio River Authority (Table 4). In addition, containment and cleanup efforts were monitored throughout the spill by Resource Protection personnel. RESULTS On September 16, 253 fish were counted in five, 150 foot segments from the spill site to IH-35 (Table 1). The first segment count was made approxi- mately one mile below the spill site where the depth of water ranged from 3 to 10 feet. There were 29 fish representing four species counted within the segment. The count included sunfish (Lepomis sp.), carp (Cyprinus carpio), gizzard shad (Dorosoma cepedianum), and white crappie (Pomoxis annularis). Organic material in the river was decomposed by the acid css which resulted in_an—increase in water Clarity compared to ambient conditions. (Many large fish were observed on the bottom of deép pools. > The pH was 1.0 (Table 4) and the fish that were counted were soft and < deteriorated. The second count was made in a shallow, riffle area approximately 500 yards downstream from the first segment. Most of the fish were caught in debris along the banks with some observed in shallow areas. One hundred five fish were counted that included sunfish, largemouth bass (Micropterus salmoides), channel catfish (Ictalurus punctatus), alligator gar Lepisosteus spatula), gizzard shad, and Rio Grande cichlids (Cichlasoma cyanoguttatum). The pH in this segment was 1.0 and the water was clear. The third count was made below Pearsall Road. This segment of the river was approximately 3 to 6 feet deep. There were 42 organisms counted representing five species that included alligator gar, American eel (Anguilla rostrata), gizzard shad, sunfish and Rio Grande cichlids. The — color of the_water_was—greenish-white as a result of lime added at-q"" Pearsall Road. The pH had-increased to 5.8 as a result of the lime. The count at segment four, made one mile upstream from IH-35, included 31 organisms representing six species such as alligator gar, gizzard shad, channel catfish, sunfish and American eel. The water was clear and a large number of dead fish were observed on the bottom. The pH of the river in this segment was 1.0. OO S The last segment count was made at the IH-35 crossing. The count totaled 46 organisms representing three species that included sunfish, gizzard Shad and American eel. Large numbers of fish were observed on the bottom. The fish were deteriorated as a result of a pH of 1.8. In summary, nine species of fish were counted in the river between the spill site and IH-35, a distance of 7.2 miles (Table 1). These species included sunfish, common carp, channel catfish, largemouth bass, gizzard shad, white crappie, alligator gar, Rio Grande cichlid and the American eel. The total count of 253 fish was then expanded to 12,829 by 0-1424-05/09/86 --- Page 5 --- multiplying the actual count of all inch classes by an expansion factor of 50.7. The pH ranged from 1.0 to 7.0 while dead fish were counted in this segment. On September 18, large numbers of dead fish were observed in the river from the spill site to Applewhite Road, a distance of 17 miles (Table 2). The pH ranged from 7.8 at the IH-35 bridge to 11.8 at Applewhite Road. The ftsh—were not soft’ and deteriorated as the ones counted on the 16th. Also, there were no dead-fish_observed on the bottom and most of the fish were floating midstream or were caught in debris along the bank. The total count within the seven segments included and 976 fish representing 15 species. The species killed were sunfish, channel catfish, largemouth bass, gizzard shad, gray redhorse (Moxostoma congestum), common carp, smallmouth buffalo (Ictiobus bubalus), Rio Grande Cichlid, Tilapia sp., alligator gar, inland silverside (Menidia beryllina), white crappie, sailfin molly (Poecilia latipinna), log perch Percina caprodes) and the American eel. ‘The total estimated number of fish killed within this section of the river was 83,449. This number was derived by expanding the total counted in all segments by the expansion factor of 85.5. Many dead fish were observed in the river within the 12 mile segment from Applewhite Road to FM 1937. On September 20, one thousand nine hundred eighty one (1,981) dead fish were counted in this segment (Table 3). The 1,981 fish counted were multiplied by an expansion factor of 42.2 resulting in an estimated number of 83,591 fish killed in this segment. The 12 species killed included the Rio Grande cichlid, channel catfish, sunfish, common carp, inland silverside, sailfish molly, gizzard shad, mosquitofish (Gambusia affinis), alligator gar, largemouth bass, gray redhorse, and American eel. On September 19, pH readings of 11.5 and 10.3 were measured within this segment at Applewhite Road and U.S. 281, respectively (Table 4). The September 16 fish kill counts from the spill site to IH-35 were not included when determining the total number of dead fish within the 29 miles of river affected. The ratio of the number of dead fish on the bottom versus those counted within this 7.2 miles was approximately 5:1. The acid deteriorated most of the fish preventing them from floating to the surface and from being counted and measured accurately. Therefore, the estimate of 12,829 dead fish in this segment was considered to be low and underestimated the full impact of the spill in this segment. The September 18 counts were expanded over this section to provide more accurate counts than those made on September 16. The counts made on September 18, between the spill site and Applewhite Road and the counts made on September 20, between Applewhite Road and FM 1937 were considered separately when determining the total number of fish killed. Since species diversity between these two sections differed, the two sections were expanded separately so that a more accurate estimate of the total number of fish killed could be determined. Several Species including tilapia, Rio Grande cichlid, American eel, sunfish and smallmouth buffalo were found exclusively or in much higher numbers at either the September 18 or September 20 sections (Tables 2 and 3). Therefore, after summing the estimated totals from both sections, the 0-142A-05/01 /86 --- Page 6 --- total kill in the 29 mile stretch of the river affected by the acid or lime was 167,040 fish. DISCUSSION When the sulfuric acid entered the river on September 14, the pH dropped from ambient levels of 7.7 to 8.1, to 1.0 at the spill site (Table 4). The acid reacted with organic matter in the river forming a black, foam- like scum. In addition, the river was dark brown in color and a slight odor of sulfuric acid was prevalent. To prevent further downstream contamination by the foam, a skim boom was placed on the upstream side of the low water crossing bridge. Monitoring of pH was started immediately by Parks and Wildlife and Water Commission personnel to locate the acid plume downstream from the spill site. The leading edge of the acid plume (pH 1.0) was located at 0230 on September 15, approximately one mile downstream from the spill site. In contrast, the pH at Pearsall Road, 2.8 miles from the spill site, was 7-8 at 0300. Discussions were held by Southern Pacific and Water Commission personnel to determine methods to stop the downstream flow of acid. Southern Pacific then attempted to construct a man-made dam across the river at the leading edge of the acid plume to divert the flow of the acid contaminated river into a gravel pit adjacent to the river. The river flow was too great to safely divert, so construction was stopped. Southern Pacific with Water Commission concurrence then made a decision to neutralize the acid with hydrated lime (Ca(OH)5). Lime was pumped into the river at approximately 0300 on September 15 at the spill site and the low water crossing bridge. At 0630 on September 15, Law Enforcement personnel observed dead fish near the man-made dam where the pH was <1l.0, much less than the 6.5 to 9.0 optimal range for fish survival. At 0930, the pH at Pearsall Road, 2.8 miles from the spill site, was also <1.0. This appeared to be the leading edge of the acid plume since the pH downstream at IH-35 was 7-8. Additional lime trucks were called in to begin liming efforts at Pearsall Road. On September 16, at 0700, the pH at IH-35 was 6.0 to 7.0 indicating slight acid contamination. At 1015, the pH measured 1.0 to 2.0 and remained Jess than 1.0 from the spill site to IH-35 for 48 hours. This low pH resulted in the mortality of 12,829 fish in this section. — Southern Pacific moved the liming operations to the downstream side of IH- 35 on the morning of September 16. Lime trucks began adding lime at a tremendous rate and the pH immediately increased to 12 to 13 within 10 minutes. The addition of lime at IH-35 was stopped at approximately midnight on September 16. The excessive liming at IH-35 reversed the condition of the river from acidic to basic. The pH of the river was then greater than the 6.5 to 9.0 optimal range and presented a hazard to aquatic organisms. The pH above IH-35 had returned to 7.8 to 8.2 by September 17. Monitoring of pH continued downstream to follow the route of the high pH plume that existed as a result of liming at IH-35. 0-142A-05/09/86 la ala --- Page 7 --- On September 17, at 0730, a pH of 13.1 was measured at Somerset Road, a distance of 2.4 miles from IH-35. On September 18, at 0900 the pH at Applewhite Road was 11.8. This high pH plume, ranging from 11.8 to i ee covered the section of the river from IH-35 to Applewhite Road, a distance of 10 miles for 48 hours. Because the pH was greater than the optimal _Occurred. Heavy rains on September 18 jn the Medina River watershed increased the flow of Leon Creek to 45 cfs. This addition of water to the Medina River helped dilute the lime below U.S. 281. The pH at U.S. 281 and FM 1937 on September 20, was 9.0 and 9.1, respectively. The section of river between Applewhite Road and FM 1937 had a high pH plume ranging from 10.7 to 11.8 from 0900 on September 18 to 1420 on September 19. A high pH plume within this 12 mile stretch of the river was responsible for killing 83,591 fish, The pH below FM 1937 was <9.0, so FM 1937 waS Considered to be the end point of the fish kill. A total of 1,078 tons of lime were used for neutralization with 525 tons placed in the river. Due to the method used to add the lime, the river became supersaturated with respect to Ca(OH) >. This resulted in a residue of unreacted lime that precipitated to thé bottom. When the Ca(OH) reacted with the sulfuric acid, a calcium sulfate crust formed over thé unreacted lime. Large deposits of unreacted lime remained in the river and would remain a continued threat to aquatic life if not removed. International Technology, Inc. (IT), a consulting firm for Southern Pacific, was hired to remove the unreacted lime residue from the river. The areas of concern were the spill site, low water crossing bridge and Pearsall Road. Approximately 25 tons of lime had been added at the spill site, 50 tons at the low water bridge and 250 tons at Pearsall Road. It was estimated that 200 tons of lime remained in the river. Removal operations began on September 24, A high capacity vacuum dredge was used to remove the lime. The dredge had an 80 foot extension boom which allowed access to most parts of the river, with minimal impact on the river bottom. The lime removed was transported to the acid contami- nated gravel pits for neutralization. When the unreacted lime was disturbed, the pH immediately increased from ambient pH levels of 7 to 8, to 14.~ To prevent further downstream contamination, a weir was constructed below the lower limit of lime contamination at the IH-35 site. The purpose of the weir was to act as a final barrier for lime flushed downstream by the dredging activities. A pH monitoring system was installed at the weir location to monitor any increase in pH caused by removal. If the pH of the water crossing the weir exceeded 9.5, a neutralization agent (H>SOg) was available to bring the pH of the river to an acceptable Tevet. hed monitored at 100 yard intervals downstream during removal operations did not exceed 8.2. Removal operations were completed by December 4, 1985. 0-142A-05/09/86 --- Page 8 --- CONCLUSION All fish in the Medina River from the spill site at FM 1604 to FM 1937, a distance of 29 miles, were killed as a result of sulfuric acid spilled during the September 14, 1985 train derailment and from over liming of the river for neutralization. For 24 hours after the spill, the pH in the river from the spill site to IH-35 was less than 1.0 which was less than the lower optimal limit of 6.5. As a result, an estimated 12,829 fish representing nine species were killed in this segment. The fish killed included sunfish, common carp, channel catfish, largemouth bass, gizzard shad, white crappie, alligator gar, Rio Grande cichlid and American eel. Over-liming at IH-35 caused a pH plume ranging from 10.8 to 13.1 to move downstream from 1015 on September 16, to 1650 on September 19. The pH was above the upper optimal limit of 9.0. As a result, 16 species were killed. The fish killed included Rio Grande cichlids, channel catfish, sunfish, common carp, inland silverside, sailfin molly, gizzard shad, mosquitofish, alligator gar, largemouth bass, gray redhorse, American eel, smallmouth buffalo, white crappie, log perch, and tilapia. Heavy runoff during the week provided additional water to the river from Leon Creek. The increase in flow diluted the lime below FM 1937 and reduced the pH to 9.0. Some fish in the river were in stress from FM 1937 to its confluence with the San Antonio River. However, the pH was within the acceptable range of 6.5 to 9.0 and a kill did not occur. As a result of the acid spill and excessive amounts of hydrated lime added to the river, 167,040 fish were killed. 0-142A-05/01 /86 --- Page 9 --- LITERATURE CITED American Fisheries Society, Water Quality Section. 1979. A review of the EPA Redbook: Quality Criteria for Water. Am. Fish. Soc. Bethesda, Maryland. Boesch, D. F. 1977. Application of numerical classification in ecological investigations of water pollution. U.S. EPA, Corvallis, Oregon. EPA-600/3-77-033. Doudoroff, Paul and Katz, M. 1950. Critical review of literature on the toxicity of industrial wastes and toxic components to fish: In Sewage and Industrial Wastes 22: 1432. Ellis, M.M. 1958. Stream pollution studies in the state of Mississippi. U.S. Department of Interior, Special Scientific Report No. 3. Harrison, A.D. 1958. The effects of sulfuric acid pollution on the biology of streams in the Transvaal, South Africa. Limnol. 13: 1603. Lewis, W.M. and Peters, C. 1956. Coal mine slag drainage. Industrial Wastes 1: 145, U.S. Environmental Protection Agency. 1976. Quality Criteria For Water. USEPA. Washington, D.C. 256 pp. 0-142A-05/01/86 --- Page 10 --- Table l. Number of fish counted and expanded numbers in Medina River from spill site to IH-35, September 16, 1985. Species Lepomis spp. Cyprinus carpio common carp Ictalurus punctatus (channel catfish Micropterus salmoides largemouth bass) Dorosoma cepedianum (gizzard shad 0-142D-05/01 /86 Inch Class 1 Z 3 Subtotal 6 8 10 12 14 LS Bmw Mr Number Counted — — = Wl Ww Mh | esdouniiolioeexe . ee — to.) Expanded Number 2281 3701 "51 6033 254 203 203 304 _ 51 1066 51 51 101 101 253 152 811 253 1217 51 1521 913 1014 _51 1978 --- Page 11 --- Table 1. (Continued) Species Inch Class Number Counted Expanded Number Pomoxis annularis 6 9 456 (white crappie) Lepisosteus spatula 24 1 51 alligator gar) 30 _l 51 2 102 Cichlasoma cyanoguttatum 6 14 710 (Rio Grande Cichlid Anguilla rostrata 20 1 51 American ee1) 24 #2 101 3 152 GRAND TOTAL 253 12,829 0-142D-05/01 /86 --- Page 12 --- Table 2. Number of fish counted, expanded numbers and monetary value of fish killed in Medina River from spill site to Applewhite Road, September 18, 1985. Species Inch Class Number Counted Expanded Number Weight (1lb./fish Lepomis spp. 2 69 5900 (sunfish) 3 403 34456 4 104 8892 5 22 1881 6 _2 171 600 51,300 Ictalurus punctatus 3 2 171 channel catfish 4 1 86 5 5 428 6 16 1368 7 if 1026 8 16 1368 9 10 855 10 20 1710 ll i 85 12 8 684 13 5 427 14 13 1112 15 1 86 16 3 256 (1.6) 18 2 171 (2.3) 20 _2 171 (3.2) 117 10,004 Micropterus salmoides 3 l 85 largemouth bass) 4 3 257 5 3 256 6 3 257 7 1 86 8 5 428 9 _d 85 Ly 1454 0-142D-05/01 /86 --- Page 13 --- Table 2. (Continued) Species Inch Class Number Counted Expanded Number Weight (lb./fish Dorosoma cepedianum 1 1 86 (gizzard shad 4 2 171 5 7 598 6 4 342 7 17 1454 8 30 2565 9 22 1881 10 26 2223 11 4 342 12 9 770 14 _2 171 (1.0) 124 10,603 Moxostoma congestum 6 4 342 (gray redhorse) 7 12 1026 8 5 427 10 4 342 ll 4 342 12 6 513 13 1 85 (.9) 14 2 171 (1.0) 18 | 86 (2.0) 39 3334 Cyprinus carpio 12 1 85 1.0 (common aan 14 1 86 ie 15 2 171 1.6 17 1 85 (2.1) 19 1 86 (3.0) 20 1 85 (4.0) 21 1 86 (4.6) 24 _t 85 (6.2) 9 769 0-142D-05/01 /86 --- Page 14 --- Table 2. (Continued) Species Inch Class Number Counted Expanded Number Ictiobus bubalus 16 (smallmouth buffalo) 22 24 Cichlasoma cyanaguttatum (Rio Grande cichlid “Sop 342 Tilapia sp ONO Ww © an nn Lepisosteus spatula 7 (alligator gar 8 20 23 (.5) 1.0) 1.1) — ~— — — Menidia beryllina (inland silverside) Mr Pomoxis annularis 6 (white crappie) 0-142D-05/01/86 --- Page 15 --- Table 2. (Continued) Species Inch Class Number Counted Expanded Number Weight (1b./fish Anguilla rostrata 14 3 256 (0.5) American eel) 15 1 86 (0.5) 18 1 85 (0.75) 19 1 86 (0.75) 20 7 598 (0.80) 24 8 684 (1.00) 21 1795 Poecilia latipinna w — © a (sailfin molly Percina caprodes 3 1 86 (log serch} 4 _2 171 3 257 SUBTOTAL 976 83,449 0-142D-05/01 /86 --- Page 16 --- Table 3. Numbers of fish counted, expanded numbers and monetary value of fish killed in Medina River from Applewhite Road to FM 1937, September 20, 1985. Species Inch Class Number Counted Expanded Number Cichlasoma cyanoguttatum 1 16 675 (Rio Grande cichlid 2 520 21944 3 242 10212 4 263 11099 5 318 13420 6 84 3545 7 2 84 445 60,979 Ictalurus punctatus 2 4 169 (channel catfish 3 25 1055 4 47 1983 5 41 1730 6 37 1561 rs 19 802 8 37 1561 9 5 211 10 21 886 11 4 169 12 9 380 13 1 42 14 6 253 15 1 42 16 7 295 (1.6) 18 5 ei (2.3) 19 3 127 (2.7) 20 6 253 (3.2) 22 1 _ 42 (4.2) 279 11,772 0-142D-05/01/86 --- Page 17 --- Table 3. (Continued) Species Lepomis spp (sunfish) Cyprinus carpio Common carp Menidia beryllina (inland silverside) Poecilia latipinna (sailfin molly Dorosoma cepedianum (gizzard shad 0-142D-05/01 /86 Inch Class obWwhr Whore anNre onaansr 10 14 Number Counted 2 39 57 23 SI ia) aS : ltt Si. HE Ore Rh RR Oe ei Ro oO Ore Om Ore ON Rr ~ Expanded Number 84 1646 2405 971 169 5275 295 42 42 42 253 380 464 1518 211 1266 42 1519 760 84 42 886 42 253 42 84 84 42 631 (1.0) (1.2) (2.5) (4.0) (6.2) (1.0) . --- Page 18 --- Table 3. (Continued) Species Inch Class Number Counted Expanded Number Gambusia affinis 1 14 591 (mosquitof ish) Lepisosteus spatula 14 2 84 alligator gar 24 1 42 (1.1) 40 1 42 (7.0) 4 168 Micropterus salmoides 4 1 42 (largemouth bass) 5 1 42 2 84 Moxostoma congestum 5 1 42 (gray redhorse) 10 1 42 2 84 Anguilla rostrata 20 1 42 (0.80) (American ee1) 24 1 42 (1.00) 2 84 SUBTOTAL 1981 83,591 0-142D-95/01 /86