Climate, Physiography, Geology, and Geomorphology of Puerto Rico

The Cordillera Central region composes about 60% of the 8711-km² island of Puerto Rico and reaches a maximum elevation of 1338 m above mean sea level. Mean annual rainfall for the island for the period 1961 to 1990 was 1600 mm (Díaz et al., 2004). Mean annual evapotranspiration varies across the island, ranging from 27% to 53% of annual rainfall in eastern watersheds (Larsen and Concepción, 1998). In contrast, the mean annual potential evapotranspiration in dry southern coastal areas is estimated to equal or exceed mean annual rainfall (Calvesbert, 1970).

The island can be subdivided into four drainage regions: north, south, east, and west, largely based upon the configuration of the Cordillera Central (Figure 2B). The northern and eastern regions receive abundant orographic rainfall associated with northeast trade winds.

Runoff from intense precipitation delivered mainly by hurricanes and other tropical disturbances transports large volumes of sediment-rich waters to the coast and shelf (Gupta, 2000). A major hurricane passes over Puerto Rico, on average, once every 10 to 20 years (Neuman et al., 1990; Scatena and Larsen, 1991). Winter frontal storms also produce substantial runoff that erodes and transports large amounts of sediment to the coast.

The combination of steep slopes, warm temperatures, high mean annual rainfall, episodic high-intensity storms, and deeply weathered soils developed on the Cretaceous volcaniclastic and intrusive rocks of the Cordillera Central results in a large number of landslides and debris flows, especially in areas that have been anthropogenically disturbed (Figure 2) (Larsen and Simon, 1993; Larsen and Torres-Sánchez, 1992, 1998). These landslides and debris flows produce large volumes of sediment that are available for transport by streams and rivers (Larsen and Santiago-Román, 2001).

Puerto Rico Coast and Shelf

The morphology of the coast and shelf of Puerto Rico varies from zones of rocky headlands at the northeast corner of the island, to alluvial plains along most of the southern and west coasts, limestone cliffs along the northwest coast, to alluvial plains with swamps and lagoons along the north coast (Warne et al., 2005). Unlike many shelf settings around the world, where the shelf morphology and sediment distribution are relict, the Puerto Rico shelf morphology and sediment distribution are in equilibrium with modern processes (Pilkey et al., 1978; Schneidermann et al., 1976). The widespread calcareous sand substrates of the shelf indicate that much of the terrigenous sediments are transported to and deposited along the shelf edge and slope (Rodríguez et al., 1992, 1998; Schneidermann et al., 1976). Bush (1991) documented that storm-induced, across-shelf transport processes dominate, and he calculated that 90% of the river sediment discharged at the coast is transported to the shelf edge and slope within a few months.

Shallow littoral currents are to the west and generally shore-parallel, and the mean tidal range along the coast is 0.34 m (Puerto Rico Oceanographic Project, 1972). Ebb tides along the north coast can carry some sediment several kilometers to the east of the river mouths; nonetheless, the majority of the sediment discharged to the north coast is deposited to the north and west of the river mouths (Bush, 1991).

Effects of Suspended Sediment and Sedimentation on Coral Reefs

Coral reefs grow within a limited range of temperature, salinity, nutrient, wave energy, and turbidity conditions. When conditions are outside one or more of these ranges, reefs become stressed (Rogers, 1977). There are a number of reef stressors, including high sedimentation rates; high water turbidity; extreme water temperatures; changes in salinity; high nutrient loads; eutrophication; physical damage by storms; overfishing; physical damage by boats, ships, and fishing gear; and atmospheric deposition of contaminants or pathogens by aeolian dust (Brown and Ogden, 1993; Glynn, 2000; Harvell et al., 1999; Shinn et al., 2000; Torres and Morelock, 2002) (Table 1). Indicators of coral reef stress include slow coral growth rate, coral bleaching, low coral-species diversity and low population density, high density of filamentous and fleshy algae, high density of sponges, and coral colonies covered by fine sediment.

Table 1.

Summary of major coral reef stressors.

Suspended sediment may contain organic matter that can serve as an energy source that promotes coral growth up to a maximum continual concentration of 10–20 m L⁻¹ or a depositional rate of 200–1000 mg cm⁻² d⁻¹ (Rogers, 1977). Chronic concentrations or depositional rates above these levels cause a reduction in coral growth rates because of smothering and reduced light levels (Hubbard, 1986; Rogers, 1990; Tomasick and Sander, 1985). Rogers (1977, 1990) proposed that (1) sediment influx of 200–1000 mg cm⁻² d⁻1 and more is incompatible with healthy reefs; (2) continual suspended-sediment concentrations of 10–20 mg L⁻¹ or more appear to be critically high; and (3) Secchi disk depths of less than 4 m probably indicate an environment in which coral reefs are stressed. In a study of coral cover and linear extension rates at five sites in southwestern Puerto Rico, Torres and Morelock (2002) documented sediment deposition rates that ranged from less than 1 mg cm⁻² d⁻¹ to approximately 10 mg cm⁻² d⁻¹. They noted that among the species studied (Montastrea annularis, Siderastrea siderea, and Porites astreoides), M. annularis cover decreased substantially where sedimentation was high, whereas the cover of S. sidereal and P. astreoides was not affected.

Although coral polyps are capable of physically removing sediment, this action requires energy that then becomes unavailable for skeletal growth or reproduction; in situations where corals are covered with a thick layer of sediment, the colony quickly dies (Rogers, 1990). Reduced transparency of the water in which corals live limits the amount of light available for the symbiotic zooxanthellae. In addition to stress caused by smothering and reduced light transmission, high rates of sediment influx and resultant deposition covers hard substrates, thereby reducing locations available for coral larvae to become established and start new colonies.

In reef zones where sedimentation rates are high, Rogers (1990) suggests that, relative to areas with less sedimentation, there is (1) lower species diversity, with less sediment-tolerant species absent; (2) greater abundance of branching forms and species resistant to sediment and to reduced light levels; (3) less live coral; (4) lower coral growth rates; and (5) an upward shift, or compression, in depth zonation to maintain sufficient light-intensity levels.

Surface-Water Discharge, Sediment Discharge, and Water Quality

Mean annual runoff to the coast for the water years 1990 to 2000 is estimated at 911 mm (Table 2) or 57% of the mean annual precipitation of 1600 mm. In a typical year, the largest daily discharge accounts for 3% to 25% of the total annual discharge (Warne et al., 2005). The proportion of the highest recorded daily water discharge relative to the mean annual total is much less than the proportion of the highest recorded daily sediment discharge relative to the mean annual total (Table 4). This difference reflects the incapacity of baseflow (low streamflow conditions) discharge, which comprises a substantial portion of the annual water discharge volume, to transport a large mass of suspended sediment. Mean annual suspended-sediment discharge from Puerto Rico into surrounding coastal waters is estimated to range from 2.7 to 9.0 million metric tonnes (t) for the water years 1990 to 2000 (Table 5). Estimated mean annual suspended-sediment yield ranges from 570 t km⁻² to 1900 t km⁻². The mean maximum daily water discharge at the nine suspended-sediment stations accounts for 23% to 58% of water discharged during a typical year (Table 4). During the year with the highest recorded daily sediment discharge (in most cases, associated with Hurricane Hortense in 1996 or Hurricane Georges in 1998), the maximum daily sediment discharge was between 44% and 90% of the total sediment discharge for that year. The highest recorded daily sediment discharge was between 1.0 and 5.6 times the mean annual discharge and 1.0 and 32 times the median annual discharge for water years 1990 to 2000 (Table 4).

The particle-size distribution of sediment in suspension at the five sediment stations with recently published data shows that 89% of the samples contained 80% or more silt and clay. These data represented runoff that ranged from 1 mm/d to 1500 mm/d (Díaz et al., 2004). Fine sediments such as these could easily be transported across the insular shelf in turbulent suspension during large storms. Although the fine sediment could reach coral reefs around the island, it is not known how much of this sediment is likely to be deposited on reefs. The high wave energy normally associated with meteorological conditions producing heavy rain, tropical disturbances, and winter cold fronts would inhibit deposition of fine sediment in near-shore areas (Schneidermann et al., 1976). Bedload sediment, characterized in only a few locations in Puerto Rico by Simon and Guzmán-Ríos (1990), has been observed in deposits discharged in the vicinity of a river mouth where it is sorted by waves and currents. Coarse material becomes incorporated into the active beach system, whereas finer sands and silts move offshore (Bush, 1991). Bed material effects are therefore likely to be limited to fringing and patch reefs within the littoral zone.

Nitrogen and phosphorus concentrations in the 24 rivers examined in this study were within the safe drinking-water standards established by the U.S. Environmental Protection Agency (2000) and the Puerto Rico Environmental Quality Board (1990). As noted by Warne et al. (2005), the concentration of dissolved solids is not highly responsive to discharge (for example, the Río Grande de Manatí station 50038100, which is typical of many rivers in Puerto Rico and elsewhere) (Díaz et al., 2004; Milliman and Syvitski, 1992; Stallard, 1995). Even though there is a slight decrease in concentrations of dissolved solids (including nutrients) with increasing discharge, it is not nearly enough to offset the increase of mass loading associated with increased discharge during floods. Most samples, however, were collected during relatively low-flow conditions and may not be representative of dissolved-solids concentrations during flood discharges. Fecal coliform and fecal streptococcus concentrations at the 24 water-quality gaging stations typically do not meet the sanitary-water standards established by the Puerto Rico Environmental Quality Board (1990) (Figure 7). Additionally, high fecal coliform concentrations are associated with eutrophic conditions that can result in a proliferation of macroalgae (Table 1). There is a slight increase in fecal coliform concentration with increasing discharge, which may be attributable to flushing of ephemeral channels and riparian areas in rural areas of the island with little to no sewage treatment (Hunter and Arbona, 1995; Larsen and Concepción, 1998). Most samples, however, were collected during relatively low-flow conditions and may not be representative of fecal concentrations during floods.

River Water and Sediment Discharge Associated with Hurricane Georges

Stormflow is the dominant process that transports sediment and nutrients from uplands to the sea. Hurricane Georges was evaluated to determine storm effects on the discharge of sediment and nutrients to the coast and shelf. On September 21, 1998, the eye of Hurricane Georges made landfall in eastern Puerto Rico (Smith et al., 2005). Hurricane-force winds and torrential rains affected the entire island until the storm moved off the western shore about 18 hours later. The Category-3 hurricane produced heavy rainfall, with 2-day rainfall totals that ranged from 10 cm to about 63 cm (Figure 6B), resulting in severe flooding and numerous landslides in the Cordillera Central (Larsen and Santiago-Román, 2001). Hurricane Georges caused the highest recorded daily discharges in nine of the 29 stream-flow-gaging stations evaluated in this study. Peak flow recurrence interval data indicate that Hurricane Georges had a highly variable impact across the island, but was especially severe in the west (Ramos-Ginés, 1999).

Precipitation and discharge estimates by island region indicate that about 2.6 billion m³ of precipitation fell on the island from September 20–25, 1998, and about 1.0 billion m³ (40%) was converted to direct runoff (Table 6, Figure 5). The precipitation estimates indicate that 6-day totals ranged from 200 mm in the east to 360 mm in the west, and the island-wide average accumulation was 300 mm. Total daily runoff for the entire island is estimated to have been 120 mm from September 20–25, 1998 (equal to 13% of the estimated mean annual runoff of 911 mm). Most of the 180-mm difference between rainfall and runoff (1.6 billion m³) would have infiltrated into temporary soil storage and long-term groundwater storage, with evapotranspiration accounting for the rest. Evapotranspiration rates in low elevation areas of the island generally are less than 5 mm/d (Harmsen et al., 2003). Groundwater levels responded relatively quickly to the heavy rains; for example, between September 20 and 30, 1998, eight wells around the island exceeded their recorded high ground-water levels (Díaz et al., 2004).

The highest daily sediment discharges caused by Hurricane Georges were more than the median annual discharges for seven of the nine daily sediment stations evaluated (Tables 4 and 7). Single-day water discharges ranged from 11% to 226% of the mean annual discharge volume, and sediment discharge was up to 19 times greater than the median annual sediment discharge. Discharge and sediment concentration rose rapidly and diminished rapidly following passage of Hurricane Georges. From 62% to 99% of the sediment discharge associated with hurricane stormflows was recorded on a single day. Maximum daily sediment yields associated with Hurricane Georges generally were proportional to runoff but spatially variable, ranging from 50 t km⁻² d⁻¹ to 6900 t km⁻² d⁻¹ (Table 7). There is, however, a consistent regional pattern: sediment yield was moderate in the north, low in the east (where rainfall and runoff were relatively low and the sediment stations are downstream from areas with extensive forest cover), and high in the south and west, where rainfall and runoff were high (Figure 5).

Table 7.

Summary of water and suspended-sediment discharge at selected stations, Puerto Rico, September 20–25, 1998. Hurricane Georges crossed the island on September 21–22, 1998. ID = identification number, m³ = cubic meter, m³ d⁻¹ = cubic meter per day, mm = millimeter.

During the period of September 20–25, 1998, a total of 2.4 million metric tonnes of suspended sediment is estimated to have been transported past the nine sediment stations, with most sediment transport (84%) occurring on the day of highest discharge. Because these nine stations represent only 19% (1622 km²) of the island area, sediment discharge to the coast may have been on the order of five times this amount, or about 12 million metric tonnes. Actual island-wide sediment discharges are likely to have been less than this, however, because of the low-relief coastal plain, which reduces sediment production and yield and provides opportunities for sediment deposition as river gradients decrease and channels widen near the coast (Figure 2B). A reasonable estimate of total sediment discharge would likely range between two and four times the cumulative amount represented by the nine sediment stations: 5–10 million metric tones (570–1150 t km⁻²). Most of the sediment was discharged from the western, southern, and northern rivers, whereas the eastern rivers had relatively little sediment discharge. In comparison, floods associated with Super-Typhoon Herb (1996) in tectonically-active Taiwan (32,260 km²) are estimated to have been 217 million metric tonnes (673 t km⁻²), according to Milliman and Kao (2005).

Nutrient discharge associated with Hurricane Georges was estimated conservatively at 1000 metric tonnes of nitrogen and 500 metric tonnes of phosphorous assuming average concentrations in streamflow (Figure 7, Table 6). The discharge of fecal material to the coast and shelf during Hurricane Georges varied among the rivers. Data are insufficient to estimate total fecal discharge volumes; however, analysis of Sea-viewing Wide Field of view Sensor (SeaWiFS) satellite imagery used for monitoring plankton and sedimentation levels in the oceans for the period of September 19, 1998, to October 15, 1998, indicates that discharge of nutrient-rich water induced a substantial increase in chlorophyll-a (planktonic algae) concentrations across and seaward of the Puerto Rico shelf (Gilbes et al., 2001; Warne et al., 2005).

River and Sediment Discharge to Coast and Shelf

Mean annual runoff from Puerto Rico is estimated at 911 mm, bearing in mind uncertainties associated with ground-water contributions to streamflow and evapotranspiration in coastal areas. Mean annual suspended-sediment discharge to surrounding coastal waters is estimated to range from 2.7 to 9.0 million metric tonnes for the water years 1990 to 2000. The relation between mean annual sediment yield (and mean annual sediment load) vs. drainage area for Puerto Rico rivers is comparable to that of other rivers of the world draining similar terrain (Figure 4). Mean annual sediment yield vs. runoff, however, indicates that Puerto Rico rivers are sensitive to the high frequency and magnitude of the tropical rainfall characteristic of this region (Larsen and Simon, 1993).

Sediment discharge from the island is highly episodic and spatially variable. Once offshore, the sediment is reworked and redistributed by waves and currents. During the water years 1984 to 1993, 80% of the annual suspended-sediment load was transported into Lago Loíza in 4 to 24 days from the Rio Gurabo branch and in 7 to 12 days from the Río Grande de Loíza branch (Gellis et al., 2006). Average sediment concentrations in slopewash measured by Gellis et al. (2006) ranged from 2000 ppm from forested areas to more than 60,000 ppm for waters flowing across bare ground at construction sites. In comparison, the average annual discharge-weighted sediment concentration was only 270 ppm from the Río Bairoa, a watershed with 6% pasture and forest (Gellis et al., 2006). Cruise and Miller (1994) estimated that runoff induced by large rainstorms in August 1988 and October 1989 produced 72% and 81%, respectively, of the sediment yields to the Río Guanajibo for the respective years. Discharge in northwestern Puerto Rico on October 21, 1989, resulted in sediment plumes that extended up to 12 km offshore (Figure 8). Suspended-sediment loads measured at the Río Rosario (USGS station 500136400, drainage area 47 km²), a tributary of the Río Guanajibo, totaled 1800 metric tonnes for the 4-day storm represented, in part, by the October 21, 1989, image and equaled about 6% of the annual total.

Figure 8.

Space shuttle photograph of Puerto Rico showing sediment plumes at the river mouths of (A) Río Guanajibo, (B) Río Grande de Añasco, (C) Río Grande de Arecibo, and (D) Río Grande de Manatí. The brown color of vegetation in eastern Puerto Rico is the result of defoliation caused by Hurricane Hugo, which struck the island on September 18, 1989. Image STS034-076-088, taken on October 21, 1989, at 12:51 GMT. Courtesy of NASA ( http://www.nasa.gov/multimedia/imagegallery/). Spacecraft elevation 179 nautical miles (332 kilometers), sun angle 32 degrees.

The episodic nature of sediment export from the island of Puerto Rico is typical of many tropical and temperate regions (Gellis et al., 2006; Gupta, 2000; Larsen and Santiago-Román, 2001; Meade et al., 1990; Sommerfield and Nittrouer, 1999). Meade et al. (1990) demonstrated that sediment is discharged during 1% (about 4 days) of the year in North American rivers and varies regionally. About 10% of annual sediment loads are discharged during 1% of the year in the coastal plain rivers of North Carolina, whereas 60% to 93% of the sediment is discharged during 1% of the year in the Pacific Coast ranges (Meade et al., 1990). The present study demonstrates that for the water year with the highest daily discharge volume, the daily discharge comprised between 8% and 66% of the total annual water discharge volume and that up to six times the mean annual sediment load is discharged in a single day. Additionally, during a large storm such as Hurricane Georges, the maximum daily sediment discharge may be many times the mean (and median) annual load (Table 7). Because sediment yields during Hurricane Georges were based upon sediment discharges at sediment stations located in downstream river reaches, rather than from sites of hillslope erosion, the sediment yields should reasonably approximate the volume of sediment that was transported from the uplands and river valleys to the coast during the storm.

In the large watersheds of mainland United States, only about 10% of the sediment eroded from uplands is transported directly to the oceans by rivers (Meade et al., 1990). Some of the eroded sediment is trapped in reservoirs, but most is stored on hillslopes, flood plains, and other parts of stream valleys. In Puerto Rico, where stream lengths are relatively short, channel gradients are high and stream valleys are steep and narrow, and where intense rainfall and high runoff are common, it is likely that well more than 10% of the sediment eroded from uplands is discharged to the coast. Mertes and Warrick (2001) indicate that the transfer of sediment to the ocean from steep, small- to moderate-size, coastal watersheds is relatively efficient because (1) sediment yields are high in comparison to sediment yields of large rivers; (2) the steep narrow flood plains do not have the capacity to store large volumes of sediment; and (3) mountainous coastal watersheds typically lack estuaries and deltas that could retain sediment. The relatively high sediment yield with respect to runoff from Puerto Rico rivers as compared to other rivers is seen in Figure 4A.

Relatively little is known about the type, frequency, and intensity of shelf processes that transport terrigenous sediment to the shelf edge (Pilkey et al., 1978; Rodríguez et al., 1992, 1998; Schneidermann et al., 1976). Bush (1991), however, documented that storm-induced, across-shelf transport processes are predominate along the north coast of Puerto Rico, and he calculated that 90% of the river sediment discharged at the coast is transported to the shelf edge and slope within a few months. Along the east coast, easterly winds and currents confine river discharge to nearshore areas, resulting in chronically turbid waters unfavorable for coral development. However, within a kilometer eastward and upcurrent, clear ocean waters are home to some of the best developed reefs on the insular shelf. At any point on the insular shelf, the average water quality reflects the interaction of waves and ocean currents with contributions of active and relict carbonate and terrestrial sediment sources.

Nutrient Discharge to the Coast and Shelf

Particulate and dissolved carbon and nitrogen are common companions to river-borne sediments and are also dominant components in mangrove forests and coastal mudflats. Nutrient-rich sediments rejuvenate flood plains and are an important energy source for the flora and fauna of coastal lagoons, but can have deleterious effects on coral reefs. When nutrients are introduced into normally oligotrophic coastal waters, the growth of photosynthetic algae is enhanced.

The blue-green ratio of backscattered sunlight detected with a radiometer can be used to estimate the concentrations of chlorophyll-a. Radiometers aboard the SeaWiFS satellite were used to quantify chlorophyll-a concentrations around Puerto Rico before, during, and after the passage of Hurricane Georges (Figure 9, Gilbes et al., 2001). The images show increased phytoplankton in the coastal areas around Puerto Rico that presumably resulted from the assimilation of nutrients that were discharged from the island during Hurricane Georges. The areas of high chlorophyll-a concentration are most intense off the west and south coasts where flood discharge was the greatest. It should be noted that humic acids are concentrated in mangrove wetland swamps. These substances absorb blue light, like chlorophyll-a. As such, some fraction of the estimated chlorophyll-a concentrations may be misclassified, especially in and around mangrove forests.

Figure 9.

Chlorophyll-a as estimated using the SeaWiFS satellite sensor just before and 4 days after Hurricane Georges crossed over the Caribbean region. Increased concentrations of chlorophyll-a around Puerto Rico after Hurricane Georges reflect an increase in phytoplankton. The increase in phytoplankton is assumed to be the result of a high discharge of nutrient–rich river water. SeaWiFS images courtesy of the University of Puerto Rico Space Information Laboratory (Gilbes et al., 2001).

Coral Reef Degradation and Land Use

Coral reef degradation is widespread in waters surrounding the island, but is generally greatest offshore from watersheds such as those in and near San Juan, where the population is high and terrestrial discharges of water and sediment are high (Figure 1). Samples with some of the highest bacteria concentrations measured at USGS streamflow stations are collected from station 50049100 (Río Piedras), which drains into San Juan Bay on the north coast (Figures 3 and 7A). Although portions of the southcentral and southwest coasts have relatively healthy reefs, disease outbreaks have been reported in these areas by Weil (2004), as described below. Watersheds that contribute runoff to these areas are not highly populated, and mean annual runoff is relatively low. Continued urbanization, however, may further threaten these areas. In 1896, approximately 21% of the island was in active cropland. In 1939, active cropland reached a high of 48% (Birdsey and Weaver, 1987). Active cropland is a major source of sediment that is entrained during storms and transported to coastal areas (Larsen and Santiago-Román, 2001). Agricultural land in Puerto Rico has been observed by the authors to be in frequent transition between active crop-land and pasture, which likely maintains high sediment yield in these areas (Gellis et al., 2006). Forest cover reached a low of 7% in the mid 1940s, but a shift from predominantly agricultural to a manufacturing economy resulted in an increase in forest cover to 33% by the 1990s. Currently, new construction is widespread and as undeveloped space in urban areas diminishes, development on steep slopes increases (Helmer et al., 2002), which contributes to an increase in sediment erosion and transport to the coast.

Extensive hillslope erosion associated with 19th and 20th century agricultural practices in the upper Río Grande de Loíza watershed is estimated to have delivered annually more than 5000 t km⁻² of hillslope-derived sediment to river channels for decades (Larsen and Santiago-Román, 2001). The erosion of hillslope soil cover was estimated to be equivalent to lowering the entire watershed by 66 cm (Brown et al., 1998). Larsen and Santiago-Román (2001) provided evidence that during the period of widespread agricultural activity, a massive quantity of sediment eroded from the hill-slopes was deposited on the lower hillslopes and in the stream valleys. Sediment in these near-channel areas provides an abundant, easily available source of material for transport by moderate to large storms. Larsen and Santiago-Román (2001) estimated that the stored sediment could sustain disturbance levels of sediment yield for a century or more. Gellis et al. (2006) reported that sediment discharge is higher than presumed pre-European background levels in the upper Río Grande de Loíza watershed, even though much of the watershed was converted from agriculture back to forested cover more than 30 years before. They proposed that after the watershed was reforested, about 30 years were required to flush the stored sediment through the system, and sediment discharges would then begin to return to predisturbance levels. These studies indicate that the island shelf, particularly seaward of river mouths, will be strongly influenced by the influx of agricultural-period–derived sediment for at least the next several decades to perhaps a century or more.

Increased turbidity along the shelf, seaward of Ponce and La Parguera (Figures 1 and 8), is believed to be the result of the conversion of adjacent upland areas from forest to agricultural and then to urban-industrial use. The increased turbidity may have resulted in a compression of coral depth zonation accompanied by changes in the relative abundance of coral species, which is directly related to individual species tolerance to sediment stress (Acevedo et al., 1989; Hubbard, 1986). Acevedo and Morelock (1988) and Acevedo et al. (1989) documented diminished coral cover, reduced species diversity, and a shift to slower-growing coral species with increasing turbidity in the Ponce area. Morelock et al. (1980) documented the decline of coral reefs in the Guayanilla and Tallaboa Bays and in the adjacent submarine canyon system on the south coast of Puerto Rico. They attributed reef decline to increased sediment discharge from the Río Yauco, Río Guayanilla, and Río Tallaboa, and from resuspension of bottom sediments by ship traffic (also note the bacteria increase, 1990–2000, Río Guayanilla, Figure 7C). According to Weil (2004), tissue necrosis affects coral colonies of Montastraea faveolata at depths of about 10 m in several reefs off the south coast. Weil (2004) also indicated that White Pox disease, Patchy Necrosis syndrome, and Dark Spots syndrome are present in coral reefs in local areas off the south and southwest coast of Puerto Rico. Increased river-sediment discharge, resulting from agriculture and construction, combined with industrial effluents discharged directly into the Mayagü ez Bay, are suspected of causing a reduction in living coral cover on the Algarrobo Reef and Escollo Rodríguez to less than 2% (Morelock et al., 1983).

Continual resuspension and transport of dredged sediments can cause reef degradation years after dredging ceases (Rogers, 1990). Dredging within the Torrecilla Lagoon in the 1950s and 1960s left behind deep anoxic pits with ammonia concentrations exceeding 15 mg L⁻¹ and biological oxygen demand exceeding 200 mg L⁻¹ (Ellis, 1976). Occasional turbulent mixing of these stagnant, nutrient-rich waters with surface waters was likely responsible for the destruction of the well-developed reef system northwest of Boca de Cangrejos (Goenaga and Cintrón, 1979). The sediments of the lagoon included discharge from sewage-treatment plants. Before dredging, the reef included extensive coral communities that extended from the sea surface down to more than 10 m. After dredging, few living corals exist at this location below a depth of about 1.5 m according to Goenaga and Cintrón (1979). With the regenerative capacity of the barrier reef gone, the average wave energy has likely increased, as has been shown in studies of the 2004 tsunami in the Indian Ocean where more damage occurred on land where natural coastal buffers (e.g., mangrove forests, reefs) had been degraded (Chatenoux and Peduzzi, 2007). Negative effects of river-derived sediment and nutrient discharge are especially pronounced in nearshore areas of the north, southwest, and west coasts. Major effects include reduced coral abundance and concomitant increased algal and sponge density and diversity. Fast-growing algae and sponges outcompete coral planulae in colonizing available substrates (Weil, 2004).

Nutrients and Coral Conditions

High nutrient and fecal concentrations in river water are a potential threat to Puerto Rico coral reefs. Nitrogen concentrations in most rivers typically range from 1.0 mg L⁻¹ to 2.5 mg L⁻¹ (Figure 7). In comparison, streams draining the relatively undisturbed forests of eastern Puerto Rico have nitrogen concentrations ranging from 0.15 mg L⁻¹ to 0.25 mg L⁻¹ (McDowell and Asbury, 1994). Phosphorus concentrations in most rivers of Puerto Rico range from 0.1 mg L⁻¹ to 0.8 mg L⁻¹, whereas in the relatively undisturbed forests of eastern Puerto Rico, total dissolved-phosphorus concentrations of 0.002 mg L⁻¹ are common (McDowell and Asbury, 1994). This limited comparison suggests that human modification of the island landscape has caused nitrogen and phosphorus concentrations to increase 10-fold. However, as a result of changes in wastewater collection and treatment brought about in response to the Clean Water Act of 1972 (33 U.S.C. 1251 et seq.), concentrations of nitrogen, phosphorus, and bacteria measured in the streams and lagoon of the San Juan metropolitan area began to fall beginning around 1985 (Webb and Gómez-Gómez, 1998). Although the concentrations of nitrogen and phosphorus are decreasing and nutrient concentrations are within drinking-water standards (U.S. Environmental Protection Agency, 2000), the overall increase in concentration of fecal material, nitrogen, and phosphorus in river water in comparison to predevelopment levels provides a competitive advantage to algae and sponges, which now dominate many Puerto Rico reefs (Goenaga and Cintrón, 1979). An outbreak of Patchy Necrosis syndrome in November 2000 was documented on a Puerto Rico reef after a 15-day period of no wind, little water movement, minimal cloud cover, and above-normal sea surface temperatures. Although Weil (2004) attributes this incident to elevated levels of fecal material from marine fauna that were observed to be deposited directly onto coral, it is possible that sewage effluent may have played a role. The month of November 2000 was one of relatively low runoff for most of the island, except for the east coast. During periods of low runoff, effluent from sewage-treatment plants and untreated sources reaches the coast with less dilution than normal, thereby contributing above-average concentrations of nutrients and other human, agricultural, and industrial waste material. Patterson et al. (2002) have suggested that untreated sewage may be a source for some coral diseases. During the decade 1990–2000, bacteria concentrations increased at a river in this part of the island (see Río Guayanilla, station 50124700, Figures 3 and 7C).