Selective oak removal does not harm water quality
Measurements before and after removal of oaks from 14% of a 250-acre watershed indicated small but not statistically significant increases in the runoff/rainfall ratio and no change in nutrient or sediment removal. Careful, selective oak removal appeared to have no harmful effect on water quality.
For the past 10 years we have measured amounts of rainfall and runoff, and concentrations of suspended sediment, major cations and anions, pH, and electrical conductivity in two watersheds at the Sierra Foothill Range Field Station. One has a perennial flowing stream, and the other, a stream that flows only in the winter. The watersheds are used for winter and spring grazing by beef cattle, and our purpose was to learn what effects selective oak removal for range improvement would have on watershed hydrology and water quality.
The uppermost portions of the watersheds were cleared of all trees between 1964 and 1966 to enhance range production. There was no further management of the watersheds until 1984 when additional, selective oak harvesting was initiated. Cutting and removal of the trees by a fire-wood contractor was done primarily during the dry summer months to minimize watershed disturbance. Nonmerchantable wood was piled and, after a few years, burned.
Water measurements began in 1980 before the selective removal of oak trees on the two watersheds. In this report, we discuss the effects on the 254.4-acre watershed with the perennial stream. Oaks were harvested on about 14% of the watershed beginning in July 1984 and ending in May 1986. Approximately 1,350 trees were removed on 37 acres.
Results and discussion
Water yleid. An annual average 27.5 inches of rainfall was measured in the watershed during the last nine years. The average consists of two heavy rainfall years (1981-82 and 82-83) and several years that were well below the station average of 29 inches a year (table 1). Runoff from the watershed varied between 3 and 30 inches and averaged 15.8 inches. The annual ratio of runoff to rainfall varied between 0.143 and 0.775 (table 1). This ratio is useful when comparing runoff among years.
There was no statistically significant difference between the average precut and postcut rainfall, runoff, or runoff/rainfall ratio for the watershed. Rainfall (34 versus 22 inches) and runoff (18 versus 14 inches) were higher in the four precut years than in the postcut years, and the ratio was slightly higher postcut (0.610) than precut (0.489). We conclude that careful removal of this small area of oaks did not have a major impact on water yield. There is a possibility that there may be a long-term increase in water yield. If the first postcut year is not included in the analysis, the average runoff/rainfall ratio is 0.665 but the difference is not statistically significant.
Researcher demonstrates the flume used to measure the amount of water flowing out of the field station's Schubert watershed.
Comparisons between similar years show an apparent increase in runoff for the same rainfall. For example, in the 1980-81 precut year, 21 inches of precipitation generated 3 inches of runoff, while in the 1987-88 postcut year, 19 inches produced 11 inches of runoff (fig. 1). The higher runoff in the postcut year can be attributed to the distribution of rainfall, particularly the “priming” effect of the November and December storms followed by heavy rainfall in January. The early winter storms refilled the depleted soil moisture storage so that the January precipitation quickly saturated the soil and produced runoff. A similar pattern occurs when precut 1983-84 is compared with postcut 1988-89. In the precut year, 29 inches of precipitation caused 14 inches of runoff, while in the postcut year, 28 inches of precipitation caused 19 inches of runoff.
The two precut years with highest rainfall (1981 through 1983) had a combined runoff/rainfall ratio of 0.66 compared with a ratio of 0.64 for the one postcut year (1985-86) when rainfall was above average. In the precut year 1980-81, runoff during every month of the rainy season was far less than rainfall (fig. 1). In the postcut year 1987-88, runoff exceeded rainfall during two months (fig. 1). This surprising result probably is due to the heavy rainfall in the three previous months. There are several springs in the watershed, and it is clear that a source of water in addition to the rain supplies the runoff. This does not appear to be due to the cutting.
Statistical tests (simple linear regressions of runoff as a function of rainfall) were run for precut and postcut years, but no significant correlation was found. (Although the R2 was ≥0.83 for both, there were too few degrees of freedom to show significance.) From examination of the annual values and storm data, the oak harvest does not appear to have increased the runoff significantly during above-average rainfall years. Although not statistically significant, the overall average ratio of runoff/rainfall was higher from the watershed after than before cutting.
Sediment yield. The amount of sediment leaving the watershed is very small and was not adversely affected by oak harvesting. Fewer tons of sediment left the watershed after cutting than before, and tons per inch of runoff decreased about 50% after cutting. Total sediment depends on sediment concentration and runoff volume. Average suspended sediment concentration was weakly correlated to the average volume of runoff (r=0.509). The suspended sediment concentration was not significantly different between precut and postcut years (table 1). The amounts of sediment leaving the watershed may be somewhat underestimated, because samples are taken at only one location and depth, and no bedload estimate is included.
Year-to-year comparisons between similar years are less clear for sediment production than for runoff. In the postcut year 1987-88, tripling the runoff tripled the total sediment load compared with the sediment load for the precut year 1980-81 (table 1). The 35% greater runoff in 1988-89 produced nearly 2 tons less sediment than the precut 1983-84 year. Timing and intensity of precipitation events are evidently more important to the total sediment load than the removal of oaks on 14% of the watershed.
TABLE 1. Annual rainfall (RF), runoff (RO), runoff/rainfall ratio (RO/RF), suspended sediment concentration (SUS), and total sediment load (SED) for Schubert watershed S2, Sierra Foothill Range Field Station, 1980-89
TABLE 2. Annual NO3-nitrogen, sum of cations (SC), chloride (CI), and suspended sediment load (SED) for the Schubert watershed S2, 1980-89
Nutrients in runoff. There were no significant differences between the precut and postcut sum of cations (calcium, magnesium, potassium, and sodium) leaving the watershed (table 2). The totals for both groups of years are small. The average annual sum of cations removed from the watershed during the precut years was 39.4 tons compared with 37.0 tons postcut. Additional cations were lost with the suspended sediment.
The chloride data are included because chloride contamination of some wells in the valley has been a concern. Small amounts of chloride are leaving the watershed, presumably from weathering of chloride-containing rocks as soils form. We did not collect sufficient precut chloride analyses to make a precut versus postcut comparison.
Nitrate was significantly higher in postcut years (table 2). The precut average nitrate loss was 0.21 ton, and the postcut average was 1.016 tons. A comparison of monthly nitrate losses in the runoff shows that runoff and nutrient concentration were the same for precut and postcut years when there was no precipitation (June through September), but during months with precipitation, more nitrate was removed from the watershed after cutting than before (fig. 2). This appears to be related to nitrate concentration in the rainwater. Average nitrate concentration in precut years from two collection points in the watershed was 0.006 mg/L and was 1.63 mg/L from the same two collection points in postcut years. Oak harvest could not affect the concentration of nitrate (NO3) in rainfall. We must conclude that the increase in nitrogen in the stream was not due to the oak harvest.
Our results from 10 years of field monitoring of precipitation, runoff, and runoff water quality indicate that a small amount of carefully controlled oak removal has little effect on runoff volume and no effect on sediment or nutrients in the runoff. The runoff/rainfall ratio was numerically but not statistically higher for the five postcut years compared to the four precut years.
Nitrate nitrogen in runoff and rainfall was significantly higher in the postcut years. This may be a result of our collection system or analytical method. It cannot be said with any certainty that the increase in nitrogen in the stream water is due to the cutting. The total nutrients lost from the watershed are small and pose no water quality hazards.