Data from 29 streamflow-gaging stations, including 9 stations with daily suspended-sediment concentration, and data from 24 water-quality stations were compiled and analyzed to investigate the potential effects of river sediment and nutrient discharges on the coral reefs of Puerto Rico. The largely mountainous watersheds of the 8,711-square-kilometer island of Puerto Rico are small, channel gradients are steep, stream valleys tend to be well-incised and narrow, and major storms tend to be intense but brief; hence flooding is rapid with peak discharges several orders of magnitude above base discharge, and flood waters recede quickly. Storm runoff transports a substantial part of fluvial suspended sediment from uplands to the coast, as indicated by sediment data from a set of nine streamflow-gaging stations representative of runoff from watersheds considered typical of conditions in Puerto Rico. For example, the highest recorded daily sediment discharge is 1 to 3.6 times the annual suspended-sediment discharge, and runoff from major storms induces sediment transport 1 to 32 times the median annual sediment load. Precipitation associated with Hurricane Georges in September 1998 is estimated to have averaged 300 millimeters across the island, which is equivalent to a volume of about 2.6 billion cubic meters. Analysis of runoff and sediment yield from Hurricane Georges indicates that more than 1.0 billion cubic meters of water and at least 2.4 million metric tonnes of sediment (and as much as 5 to 10 million metric tonnes), were discharged to the coast and shelf as a result of this major storm. Because of their relatively small size, dams and reservoirs of Puerto Rico have relatively little effect on total discharge of water and sediment to the coastal marine waters during major storms. The presence of reservoirs, however, may be detrimental to coral reefs for two reasons: (1) coarse sediments deposited in the reservoir can be replaced by finer sediments scoured, if available, from the river channels and flood plains below the dam; and (2) the loads of phosphorus and ammonia reaching the coastal waters may increase as organic matter decomposes in the anoxic bottom waters of the reservoir. Rainfall, water discharge, sediment discharge, and sediment yield vary across the island. Mean annual runoff for the island is estimated to be 910 millimeters, about 57 percent of mean annual precipitation (1,600 millimeters). 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. Hydrologic and sediment data associated with Hurricane Georges indicate that sediment yield is generally proportional to the depth of storm runoff. Discharge and sediment-concentration data indicate that during this storm, river water and sediment that discharged into the marine environment generally formed hypopycnal plumes (buoyant suspension layers). Generally, hyperpycnal (density) plumes can develop in areas with high discharges and sediment concentrations. Both hypopycnal and hyperpycnal plumes distribute suspended sediment over broad areas of the Puerto Rico shelf and shelf slope. Comparison of long-term suspended-sediment discharge and watershed characteristics for Puerto Rico with those of other river systems around the world indicates that Puerto Rico rivers are similar to temperate and tropical upland river systems.

The Revised Universal Soil Loss Equation (RUSLE) was used in conjunction with a Geographic Information System to determine the influence of land use and other environmental factors on soil erosion in the Guadiana watershed in Puerto Rico. Mean annual erosion, suspended sediment discharge, and the rainfall-erosion factor of the RUSLE increased with annual rainfall. Median soil erosion rates varied among the seven land uses: bare soil (534 Mg ha-1 yr-1), open canopy forest (26 Mg ha-1 yr-1), agriculture (22 Mgha-1 yr -1), pasture (17 Mg ha -1 yr -1), less dense urban (15 Mg ha-1 yr -1), closed canopy forest (7 Mg ha -1 yr -1), and dense urban (1 Mg ha-1 yr -1). The differences between open canopy forest, agriculture, pasture, and less dense urban were not significantly different but median values for open canopy forests were slightly greater because they occurred on steep slopes. The five-year average sediment delivery ratio for the basin was 0.17, which is comparable to delivery ratios estimated for watersheds of similar size. Simulations of different land use configurations indicate that reforestation of 5% of the watershed with the highest erosion rates would decrease basin wide erosion by 20%. If the entire watershed was reforested, soil erosion would be reduced by 37%.