Effects of Fire Severity on Nitrate Mobilization in Watersheds Subject to Chronic Air Pollution

The setting. Atmospheric deposition of nitrogen oxides and ammonium to land surfaces is the most extreme yet recorded in the United States where the chronic urban air pollution of the South Coast Air Basin meets the San Gabriel Mountains in southern California. There the deposition and flux of nitrogen through the chaparral, which amounts to 1.7 keq ha^-1 yr^-1, induces stream-water nitrate (NO₃^-) concentrations that are as much as three orders of magnitude greater than where air pollution is minimal. Yet less than 10% of the deposited nitrogen appears in streamflow. In the absence of fire this accumulation substantially enriches the ecosystem. However, destructive wildfires that burn across tens of thousands of hectares recurrently sweep the native chaparral in the San Gabriels. Subsequent soil erosion, which is massive, combined with potentially rapid nitrification in soils and sediments may largely mobilize the accumulated nitrogen. Debris-laden flows produced by even modest storms could be heavily polluted with NO₃^- and contribute to the existing NO₃^- pollution in the aquifer of the main San Gabriel Basin, an important local source of water for Los Angeles County. If rates of soil organic matter oxidation and erosion are related to the intensity or duration of heating, then intervention in the wildfire regime by prescribed burning when weather, fuel moisture, and fuel accumulation are moderate could reduce water quality impacts by limiting the severity of subsequent wildfires. We tested these possibilities in a major watershed project at the San Dimas Experimental Forest that involved cooperation among the USDA Forest Service, California Department of Forestry and Fire Protection, and Los Angeles County Fire Department and Department of Public Works.

Severe fires were ignited in two 16-hectare watersheds after felling of the Ceanothus chaparral. The subsequent NO₃^- and ammonium (NH₄⁺) fluxes in stream water were then compared to those from two watersheds burned by more-moderate fires in undisturbed vegetation and to those from two unburned watersheds. We also examined the effect of fire severity on temporal trends in soil inorganic nitrogen.

Experimental Fires. Prescribed fires were ignited by helitorch in 24-year-old Ceanothus crassifolius - Adenostoma fasciculatum chaparral on 29 and 30 October 1984 in Bell and West Fork of San Dimas canyons. Fires burned with runs of 100 to 500 m length up the slope. The experimental fires were explosive; flames 15 to 30 meters in length burned in all four canyons, and radiant heat could be felt by observers at a distance of over one-half kilometer. Radiant and sensible heat flux from the fires in standing vegetation was estimated to be 5.1x10 11 Joules/hectare. That from the fires in felled vegetation was approximately three times greater, 1.6x10 12 Joules/hectare. The felled vegetation burned for a noticeably longer duration and was completely incinerated with less latent heat loss. Soil surface temperatures under the felled chaparral peaked at 780 C 3 minutes after combustion began; a peak of 265 C was reached at 2 cm depth in the soil after 7 minutes.

Sediment Flux and Streamflow. Even small storms during the succeeding winter produced debris-laden flows in the burned canyons, entraining sediment at up to 59% by weight. Sedimentation following the moderate-intensity fires was three-eighths as great as that from the severely burned watersheds; that from the unburned watersheds was negligible. The burned watersheds were coursed by powerful debris flows during a thunderstorm on 19 December 1984. Those flows in the severely burned canyons peaked at more than 330 L s^-1 ha^-1; concurrent flow in the unburned watersheds peaked at less than 1.1 L s^-1 ha^-1. Annual water yield from the severely burned watersheds was on average four times greater than after the moderate-intensity fires and 14 times greater than from the unburned watersheds (Table I).

Stream-water NH₄⁺ and NO₃^-. Flux of dissolved NH₄⁺ in stream water from burned canyons was dominated by ash-laden flows initiated by an early storm. Peak NH₄⁺ concentrations at San Dimas Canyon were observed during that event: 0.25 meq/L from the moderate-intensity fire and 0.75 meq/L from the severe fire. Subsequent NH₄⁺ concentrations were generally 1 to 2 orders of magnitude lower. In all cases NH₄⁺ comprised less than one-eighth the annual flux of dissolved inorganic nitrogen (Table I).

Stream-water NO₃^- concentrations responded to the passage of even small storms, rising rapidly with increasing discharge and declining slowly during recessional flows. NO3 - concentration in the large debris flows from the severely burned watersheds reached 1.12 meq/L, exceeding the Federal water quality standard (0.73 meq/L). The effect on NO₃^- yield of the intense rain on 19 December was qualitatively different from other storm periods in that it caused large quantities of soil to be suspended and removed rather than leached.

Both stream-water NO₃^- concentration and flux reflected the severity of burning. The daily-mean volume-weighted NO₃^- concentrations after severe burning were 1.7 times those after the moderate-intensity fires (with 95% confidence interval for that ratio of 1.3, 2.3). Moderate fires produced concentrations 3.0 times those of the unburned controls (confidence interval of 2.0, 5.2). The response of NO₃^- flux to fire severity was even more striking: annual NO₃^- loss from the severely burned watersheds, averaging 1.2 keq/ha, was approximately 7 times greater than after moderate-intensity burning and 40 times greater than from the unburned watersheds (Table I).

Soil nitrogen response. Concentrations of mineral nitrogen in soils were also altered according to the severity of the fire treatments. The NH₄⁺ concentration in surface soils rose during the severe fire but not as markedly as during the more moderate heating of the fires in standing vegetation. Because of the 3-fold greater energy release of the severe fires, soils in that treatment were undoubtedly subjected to the greatest heating below the immediate soil surface and the greatest rates of volatilization, which would explain their lower NH₄⁺ concentration. We also observed that soil NO₃^- concentrations to 5 cm depth declined by an average of 40% during burning.

During the ensuing winter and spring, NH₄⁺ concentrations were consistently highest in soils that had been subject to severe heating and lowest in soils of the unburned watershed. Differences in accumulated soil NO₃^- with fire severity were also apparent after a succession of winter storms when average soil NO₃^- concentration was greatest in soils subjected to severe fire (1.20 meq/kg of soil), intermediate in those from the moderate-intensity fire (0.79), and least in the unburned soils (0.49). A relative acceleration of nitrogen mineralization and nitrification in the severely-burned soils was a likely cause of early differences in mineral nitrogen between the fire treatments.

Watershed NO₃^- loss must have been partly mediated by the soil nitrification response to fire severity and not just the response to water yield and sedimentation alone. The severe fire treatment produced on average a 4-fold increase in water yield over the moderate-intensity fires yet caused a 7-fold increase in NO₃^- yield due to concurrent increases in stream NO₃^- concentration. These concentration increases did not result from the higher water yields even though high concentrations are generally correlated with high flows; at comparable flow rates NO₃^- yield was consistently higher from the severely burned watersheds than from the more moderate treatment.

Role of Atmospheric Deposition. It remains to ask whether the high rates of NO₃^- flux in stream water were a result of high rates of atmospheric deposition. A direct test is not available to us since our experimental design was not replicated in regions of low deposition. But we can make some inference regarding the high NO₃^- flux by examining three alternative explanations: that it results from the postfire environment alone; that symbiotic N₂ fixation associated with Ceanothus provides a primary source of nitrogen for nitrification and NO₃^- loss; and that the intense storm in late December entrained a quantity of NO₃^- that was unusually large for postfire situations.

High NO₃^- concentrations and rates of flux are apparently not a universal characteristic of fires in chaparral watersheds. Peak concentrations after burning in Arizona chaparral, 0.14 meq/L, were only one-tenth as great as measured here (Longstreth, D. J.; Patten, D. T. Amer. Midland Natur. 1975, 93, 25-34.). The Arizona watersheds studied were remote from any urban air basin and unlikely to have had chronically high rates of dry deposition as occur at San Dimas.

NO₃^- flux at San Dimas was also unusually high in comparison with that after fires in some coniferous forests, where nitrification is typically suppressed and recorded NO₃^- concentrations in stream water have remained below 0.04 meq/L (Tiedemann, A. R., et al., J. Environ. Qual. 1978, 7, 580-588; Fredriksen, R. L., In: Forest Land Uses and Stream Environment; Oregon State University: Corvallis, OR, 1971; pp 125-137).

Probably any source of nitrogen which is cycled through the ecosystem may equally contribute to nitrogen loss after fire. Yet although symbiotic N₂ fixation may enrich the ecosystem it is apparently not sufficient to saturate the ecosystem and drive a large NO₃^- flux in stream water in the absence of fire. This was evident in our earlier survey of unburned watersheds in southern California (Riggan, P. J., et al., Environ. Sci. Technol. 198519, 781-789.), where stream-water NO₃^- concentration at high flow was exceptionally low in regions of minimal air pollution, whether or not extensive stands of Ceanothus were present (as in portions of the western Santa Monica Mountains ). The uniformly high NO₃^- concentration in waters from unburned watersheds of the front range of the San Gabriel Mountains, as much as three orders of magnitude greater than in areas subject to low levels of air pollution, reflects in these ecosystems both a dominance of atmospheric deposition in the nitrogen cycle and a developing nitrogen saturation, or inability to sequester additional nitrogen. This is not surprising since the throughfall nitrogen flux at San Dimas, which is dominated by dry deposition to the plant canopy, is 2.3 times the rate of nitrogen accretion in the biomass of mature Ceanothus stands there (Riggan, P. J., et al., Environ. Sci. Technol. 198519, 781-789.). The deposition alone represents a very high rate of ecosystem loading, regardless of what the nitrogen fixation rates might be.

We also reject the possibility that debris-laden flows caused by unusually intense precipitation in the year of our experiment caused abnormally high rates of NO₃^- flux. Although debris flows dominated sediment movement and mobilized NO₃^- and sediment from across the burned watersheds, NO₃^- flux was dominated by inter-storm flow, and thus NO₃^- and sediment losses were not strongly coupled.

We do expect that more sustained winter precipitation would have accelerated NO₃^- loss by expanding the volume of soils contributing water to streamflow and by lowering residence time of waters and the chance for biological uptake of NO₃^-. NO₃^- yield was apparently limited by the amount of water percolating through the soil since stream NO₃^- concentrations were consistently high during peak streamflow and soil NO₃^- was by no means exhausted by leaching during winter rains. Precipitation during our observations was not unusual: annual rainfall was in the 50th percentile of a 60-year record at San Dimas.

We infer from these arguments that the very large postfire NO₃^- losses in stream water we observed at San Dimas were most likely a result of the chronic nitrogen enrichment of the ecosystem by atmospheric deposition.

Environmental Implications. By interposing young age classes in the expanse of older, even-aged chaparral, prescribed fire may interrupt the succession of extensive, high-intensity wildfires that recurrently sweep the San Gabriel Mountains . Results reported here show that moderation of this fire regime, by reducing the magnitude of postfire debris flows and the concentrations of NH₄⁺ and NO₃^- within soils and waters, would partly mitigate water pollution derived from chronic atmospheric deposition of nitrogen oxides and ammonium. Postfire waters, especially those at high flow, should also be managed downstream to avoid additions of nitrate to the aquifer of the eastern Main San Gabriel Basin , where nitrate concentrations now commonly exceed the Federal standard for drinking water (Department of Water Resources, Southern District, State of California, 1977, Nitrates in Ground Water in the Los Angeles Drainage Province).

Table I. Fluxes of water, nitrate, and ammonium and volume-weighted NO₃^- concentration in stream water during the hydrologic year beginning 1 October 1984.

For more information, please see the following:

Riggan, P. J. , R. N. Lockwood, and E. N. Lopez. 1985. Deposition and processing of airborne nitrogen pollutants in Mediterranean-type ecosystems of southern California . Environmental Science and Technology19(9):781-796.

Riggan, P. J. , F. H. Weirich, L. F. DeBano, P. M. Jacks, R. N. Lockwood, C. Colver, and J. A. Brass. 1994. Effects of fire severity on nitrate mobilization in watersheds subject to chronic atmospheric deposition. Environmental Science and Technology28(3):369-375.

Glossary:

Nitrification— The biologically mediated production of nitrate from ammonium.

Mineralization— The production of ammonium from nitrogen-containing organic compounds.

Hectare ( ha )— The metric unit of measure for land area; approximately 2.5 acres.

keq— Kiloequivalents of nitrogen; 1 keq of nitrogen weighs 14 kilograms.