The Texas Intensive Silviculture Study (TexIS)

Introduction

            Silvicultural practices have changed in the past twenty years, as the third forest in the south, a forest consisting of pine plantations, is harvested.  Contemporary silvicultural practices involve less slash disposal due to better wood utilization and more homogenous stands.  However, the use of silvicultural chemicals and fertilizers has increased in recent years, as has mechanical amelioration of soil problems through practices such as ripping and bedding.  In addition, Best Management Practices (BMPs) which minimize the potential effects of silvicultural practices on water quality have been developed and are implemented on well over 90% of harvest operations.  The Texas Intensive Silviculture study was initiated to examine he effects of these changes in silvicultural practices on surface water quality with the addition of BMPs.   

            Texas A&M University instrumented nine small (~2.5 ha) watersheds in 1980 to examine the water quality effects of clearcutting second growth mixed pine-hardwood forests and establishing loblolly pine plantations using two different mechanical site-preparation methods, rollerchopping and burning and shearing, windrowing and burning.  The results from this study indicated that rollerchopping had no measurable impact on sediment and nutrient export, while shearing resulted in an average first-year sediment loss rate of 2,937 kg/ha.

Study Design and Sampling Methodology 

            The primary objective of the Texas Intensive Silviculture Study (TexIS) was to evaluate the effects of contemporary intensive silvicultural practices with BMPs on water quality.  To accomplish this objective, these same nine small watersheds were instrumented in 1999 to examine the water quality effects of clearcutting these plantations and reestablishing loblolly pine plantations.  Two intensities of site-preparation were employed.  Conventional site-preparation included a site-preparation herbicide application, followed by machine planting, and a release herbicide application.  Intensive site-preparation included everything in the conventional method, but added a ripping operation, fertilization, and an additional release herbicide application in year two (Figure 1).  Four large (70-135 ha) watersheds were also instrumented in order to examine the water quality response from stand-sized management units.  Small watersheds were nested in the large watersheds.  Also, logging sets and forest roads were included in the large watersheds.  One large watershed served as a control, while one received the conventional site-preparation, one received the intensive site-preparation, and one with a 5 year old stand received a herbicide release and fertilization. 

            Water quantity was measured on the small watersheds with 3-foot H-flumes and programmable dataloggers.  Concrete control structures were constructed on three of the large watersheds, while flow was gauged through a 5-foot culvert on the fourth (Figure 2). Stage-discharge rating curves were developed for the four large watersheds and data were stored on programmable dataloggers.  Isco pumping samplers automatically collected storm runoff samples from the streams.  Water samples were analyzed for sediment, nutrients, metals, and herbicides by a contract laboratory according to established Environmental Protection Agency methods.  Quality assurance and quality control procedures were developed, implemented, and third-party audited during the study. 

Figure 1.  Large watershed 4 following harvest, site-preparation, and planting, late December, 2002.

Figure 2.  Stream gauging station on large watershed 2 (conventional treatment) during a storm-runoff event.

            In addition to the water quality aspects to the TexIS study, the effects of silvicultural practices on soil compaction were measured.  Soil bulk density was measured at three depths (0-4”, 4-8”, 8-12”) before treatment, after harvest, and again after site-preparation.  Soil compaction was also measured on logging sets and skid trails since these areas receive the greatest traffic and are associated with higher bulk density values, although these areas typically occupy a relatively small percentage of the treatment area.  Percent volumetric soil moisture content was measured on theses watersheds as well, so that the treatment effects on soil moisture could be correlated with any observed differences in storm runoff.

            Another area of concern in forestry today is the impact of forest roads on stream sedimentation.  Numerous studies have indicated that forest roads result in more stream sedimentation than silvicultural practices.  Therefore, in order to evaluate the potential road effects, sediment movement from the road prism was monitored on nine road segments of uniform length (100’) on three slope classes and three traffic intensities was measured.   

Results

            The following conclusions were drawn from the first year post-stand establishment data from this study:

            1.  Concentrations and losses of nutrients and sediment from these forested watersheds were generally low during the pre-treatment phase.  However, large storm events like Tropical Storm Allison can result in sediment loss rates greater than first year annual rates observed after clearcutting and rollerchopping.  These watersheds have a high potential for natural, geologic erosion. 

            2.  Storm runoff increased significantly following harvest on all six treated small watersheds, but not on the two treated large watersheds (Table 1, Figure 3).  The small, headwater watersheds generally had a steeper channel, less storativity, and more circular basin shapes.    The difference in runoff between the treated and control watersheds was greatest for storms occurring during the growing season, indicating that the increase in runoff from the six clearcut small watersheds can be attributed to a decrease in evapotranspiration due to harvest.

            3.  Following harvest and stand establishment, greater sediment loss rates were observed on the intensive large watershed than on the conventional (Table 1).  The greatest small watershed annual sediment loss rate was observed on SW6.  This loss rate of 564 kg/ha was still over 5 times lower than the rate observed in 1981 on the roller-chopped watersheds.  Also, a statistically significant increase in first-year sediment loss was only measured on 3 of the 6 clearcut watersheds.  Contemporary BMPs, especially the use of SMZs, can at least partially account for this reduction.

            4.  Nitrogen and phosphorus export significantly increased following fertilization on the intensive watersheds.  However, the total annual loss was only a small fraction of the applied nutrient (Table 1).  It is unlikely that the maximum measured first-year phosphorus loss rate of 0.77 kg/ha would have any biological significance to these streams.  Elevated loss rates were only observed for the first few storms following the December application.  Furthermore, the loss rates while statistically significant are not likely to have any negative impact on stream biota.  Loss rates were still much lower than rainfall inputs. 

Table 1.  Small watershed mean storm runoff (cm) stormflow, sediment, and nutrient loss rates (kg/ha) for the first full year after stand establishment (2003). 

Figure 3.  Sediment loss comparisons (kg/ha/yr) between the current study with BMPs and the pre-BMP 1980 study.

            5.  Fertilization of the 5 year old stand resulted in no measurable increase in nutrient losses. 

            6.  In-stream Imazapyr concentrations peaked at around 39 ppb in the first storm following harvest, and fell below the 1 ppb detection limit within about 6 months (Figure 4).  Hexazinone peaked at around 9 ppb and were below the detection limit within about 9 months.  The difference in peak concentrations can be attributed to the fact that the Hexazinone was applied in bands, where the Imazapyr application was aerial broadcast, meaning that less Hexazinone was applied to the watersheds and the drift potential was negligible.

            7.  These herbicide concentrations were orders of magnitude below the published LC₅₀ values and the herbicide application is not likely to have any negative impact on stream biota.  Herbicide losses represent only a small percentage of the total applied active ingredient, meaning that most of the herbicide was taken up by plants and soil and/or volatilized and broken down.

Figure 4.  Maximum storm runoff imazapyr and hexazinone concentrations from SW6.

            8.  Clearcut harvesting did not significantly increase soil bulk density on these watersheds (Figure 5).  Thinning the SMZ also did not result in greater soil compaction. 

However, logging sets and skid trails did have significantly greater bulk densities following harvest, with mean large watershed surface bulk densities increasing from 0.91-0.96 g/cm³ before harvest to 1.08-1.16 g/cm³ on sets and skid trails after harvest.  Site-preparation did not significantly impact soil compaction.  Based on soil infiltration studies conducted during the 1980 study, it is unlikely that these bulk densities would result in decreased infiltration and thus increased overland flow. 

Figure 5.  Mean surface (0-4”) bulk density for large watersheds after harvest.

            9.  Sediment movement from the forest road prisms varied greatly, with values ranging from 1,200 kg/ha on grassy roads with low slope and traffic, to over 14,000 kg/ha on steeper slopes with higher traffic (Figure 6).   These sediment movement rates are much higher than loss rated measured from the watersheds following harvest and site-preparation.  However, not all of this sediment would necessarily have moved into the streams.  When good road BMPs are implemented, this sediment is deposited well above stream crossings, and the actual sediment delivery to the streams is significantly lower.  Additional research is needed to quantify road sediment delivery to streams in the Western Gulf Coastal Plain.

Figure 6.  Annual sediment movement from road prisms for each road segment.

Summary and Conclusion

            Silvicultural activities generally have only a small, short lived impact on water quality, especially when compared with other land uses like agriculture.  The use of BMPs helps mitigate possible water quality effects.  The use of SMZs in particular helps stabilize stream channels and prevents direct application of silvicultural chemicals to the stream.  Harvesting practices that redistribute logging slash into erosion sensitive areas and minimizes the number of skid trails helps reduce soil movement.  Site-preparation practices that minimize the amount of bare soil on the watershed also can help minimize soil detachment and movement.  Finally, ripping and bedding must be done on the contour.  Rips and beds that are not on contour can serve as preferential surface flow paths, channeling water downslope forming rills and gulleys that can wash through SMZs providing direct conduits for sediment plumes to enter streams.  The results from this study indicate that first-year water quality effects of contemporary silvicultural practices with BMPs are significantly lower than those from harvesting and site-preparation without BMPs in the early 1980s.   Additional analysis of water quality data from the second and third years following harvest will be evaluated to determine the amount of time that these effects persist.