Dr. Richard Grippo

                       Problem/Need Statement

 The Federal Water Pollution Control Act of 1972, commonly known as the Clean Water Act (CWA), recognized Non-Point Source (NPS) pollution and called on the States to develop and implement appropriate water quality management plans.  Section 319 of the CWA directs the States to develop water quality management plans identifying Best Management Practices (BMPs) to reduce water pollution from each NPS category, including silviculture.  Since then, State foresters have been working closely with State water quality agencies and the Environmental Protection Agency to minimize sources of non-point water pollution.  For the last three decades, the forestry community, including State forestry agencies, has advocated the proper use and importance of BMPs to minimize the environmental impact to water quality associated with silvicultural activities (i.e. Ursic 1979, McClurkin et al. 1985).  Studies have shown that poor logging and skid road design and inconsistent maintenance of streamside management zones are a direct cause of sediment export to streams and other water bodies during intensive silvicultural operations (e.g. Lynch and Corbett 1990, Ursic 1991).  The retention of riparian buffers also mitigates changes in water quantity and quality during timber harvest (Wynn et al 2000). 

While much attention has been devoted to evaluating biological integrity in streams exposed to NPS from agriculture (Carline and Lieb 1996; Johnson,Ward & Grippo 2000), gravel mining (Brown & Lyttle 1992)  stream channelization (Mauney & Harp 1979) and even urban sources (McMahon and Cuffney 2000), very little work has been completed  that used bioassessment to evaluate the efficacy of silviculture BMPs.  such studies in Florida and South Carolina indicated that standard BMPs are sufficient to prevent changes in aquatic benthic macroinvertebrates (FDEP 1997, Adams et al 1995).  However, much of the topography, soil types and silviculture practices in Florida and South Carolina are unique to those states and may not apply to other areas of the country, including Arkansas.  Thus a study investigating the efficacy of silviculture BMPs in Arkansas is warranted.

 To date, most State forestry agencies have centered their efforts on information and education.  Many State forestry agencies, particularly in the South, have entered into partnership with forest industry, academia and Cooperative Extension, and established BMP standards and integrated water quality concerns into the Sustainable Forestry Initiative and associated programs such as Master Logger.  Periodic compliance monitoring is usually undertaken to develop a measure of the degree of implementation regarding these standards.  Findings from such implementation monitoring are then incorporated into subsequent revisions of State BMPs and educational programs.

The Environmental Protection Agency issued guidance on August 19, 1987, directing States to design, implement and monitor the effectiveness of established BMP protocols.  While compliance monitoring has provided a measure of BMP implementation across the Southern Region, the question of BMP effectiveness has remained largely unanswered.  Florida and South Carolina have initiated studies to develop a proven, field-tested response.

In 1994, the Southern Group of State Foresters (SGSF) appointed a task force to develop recommendations for a more consistent approach to BMP monitoring in the Southern Region.  Recommendations submitted by the task force were presented to the SGSF in 1997.  The task force recommended that individual practices, category of practices and overall site BMP implementation be evaluated and reported.  The task force further recommended that categories and overall scores be expressed as a simple percentage of all applicable practices.

However, the task force also recommended that significant risk to water quality, attributable to non-implementation for a specific BMP or category of BMPs, be assigned and documented.  Significant risk should be considered as a situation or set of conditions that can be remedied or otherwise mitigated.  The task force suggested that failure to implement BMPs that could result in adverse impact on productivity, sustainability, use or other site values should not be considered a significant risk in the context of implementation monitoring.  Significant risk should be directly and exclusively related to water quality impairment.

Documenting the occurrence of significant risk serves a number of useful and practical purposes, including enhancing the credibility and integrity, and establishing a link with effectiveness monitoring components, of the State’s silvicultural water quality program.  Finally, providing forest landowners with an objective risk assessment may also protect the landowner and/or forestry vendor from possible enforcement action and subsequent litigation and liability.

General Project Description:

Recognizing that BMPs can reduce non-point source pollution for silvicultural activities, guidelines were adopted as an implementation plan and first published in 1982 by the Arkansas Forestry Commission (AFC).  These guidelines were updated and republished in various forms through 1996.  That year, following a USDA Forest Service review requested by the AFC, a plan was adopted by the AFC Commissioners to survey the implementation of BMPs throughout the state on a biannual basis.  An education program designed to make loggers and landowners aware of the need to use BMPs was initiated at the same time.

The AFC completed its second statewide forest BMP implementation survey in July 1999.  The AFC monitored 273 tracts randomly selected through computer modeling.  The survey found an overall statewide BMP implementation rate of 80 percent.  BMP implementation was lowest among non-industrial private forest landowners, one of four categories of forest ownership recognized in the survey.  Six categories of forest activities were also evaluated.  Forest road construction and maintenance followed by harvesting received the lowest implementation ratings. 

The current BMP compliance monitoring process used by the AFC does not include an environmental risk assessment, nor has the AFC monitored BMP effectiveness to determine if there are adjustments necessary for current BMPs.  Also, in only one forest activity category, streamside management zones, is consideration given to water quality impairment and stream sedimentation.  Based on the 1999 survey results, a tentative determination could be made that water quality impairment occurred on 3-6 percent of the sites evaluated.  The survey does not address actual water quality impacts due to silvicultural treatments.  This consideration is an unknown factor given the lack of risk assessment and effectiveness monitoring data associated with the AFC’s water quality program.  In the absence of the above data a study of silvicultural BMP effectiveness in Arkansas is proposed.

Project Outcomes

This project would generate data that would used to 1) test for changes in benthic macroinvertebrate assemblage over time above and below silviculture sites, 2) test for gross changes in measured stream physico-chemical and habitat variables above and below silviculture sites, 3) provide a decision on whether current BMP practices are suitable for preventing NPS pollution from arising from silviculture sites and 4) provide input into evaluations and decisions associated with development of Total Maximum Daily Load (TMDL) in watersheds within the tested ecoregions.   These BMP assessments could be clustered with other TMDL evaluations within these watersheds to significantly reduce the cost of developing a watershed TMDL for sediments and other suspended solids .

Bioassessment Monitoring of Water Quality

Based on the success experienced in Florida, we propose to implement a BMP effectiveness monitoring study utilizing a bioassessment methodology as the principle measure of water quality and stream ecosystem health.  This follows the recommendation by the AFC, in collaboration with the Arkansas Forestry Association, Arkansas Department of Environmental Quality, Arkansas Soil and Water Conservation Commission, Arkansas Cooperative Extension Service and the University of Arkansas at Monticello.  This approach is necessary because forestry non-point source discharges are transient in nature and driven by random storm events that are difficult to account for with traditional periodic water sampling.   The bioassessments will consist of an aquatic habitat assessment and stream sampling for benthic macroinvertebrates to allow comparisons of stream ecological condition before and after silvicultural BMPs have been implemented.  The bioassessment recognizes that resident aquatic biota in a stream function as a continual natural monitor of water quality, habitat availability and ecosystem health.  Benthic macroinvertebrates in particular are sensitive to the effects of episodic and cumulative pollution as well as physical habitat alteration.  Their presence, or absence, and abundance are representative of water quality and the overall ecological integrity of the stream over time.  For this reason, bioassessment appears to be a practical and useful method for monitoring non-point source pollution and evaluating the effectiveness of Arkansas’ silvicultural BMPs.

Study Design

 Eight silviculture sites, three each in the Ozark and Ouachita regions, and two in the southern coastal plain ecoregions, will be examined.    The southern coastal plain will be examined more cursorily because most streams in this ecoregion are low gradient, silt-bottomed waterways that carry a high sediment load.  Thus, biological assemblages are already likely to be silt-tolerant  (Cobb et al. 1992) and will probably not significantly respond to any but the most egregious cases of poor BMP implementation.   Specific sites will be determined jointly by the principal investigator and the AR Forestry Commission using the following criteria 

1)      sites will be located adjacent to a perennial or mostly perennial, larger second or smaller third order stream and representative of the higher slope and soil erodibility conditions associated with silvicultural treatments used in that ecoregion. 

2)      selected sites will be located on streams which appear to be of  sufficient high water quality and habitat characteristics to support benthic macroinvertebrate assemblages which are intolerant of elevated sediment conditions. 

3)      sites will have no or minimal equipment traffic traversing the stream during logging and BMP implementation. 

4)      Sites will have upstream and downstream riffles and pools which are of similar width and depth.  This will increase the likelihood that uncontrollable climatic conditions (drought, floods) will affect upstream and downstream sites equally. 

5)      total area of timber harvests should be within the range of 75 to 200 acres (typical range of commercial silviculture in AR). 

6)      total length of stream bank bordered by the harvest should be approximately equal whether the harvest is on one side or both sides of the stream.  Harvests on one side only may make it easier to meet criterion 2).

7)      sites containing ephemeral streams within the harvest area should be avoided, especially if no stream management zone (SMZ) for the ephemeral stream is established during harvest.

8) control and impact sites should be restricted to the same  stream reach and located in similar habitat units.

 Data collection and analysis will be conducted to allow an MBACI analysis (Multiple Control, Before-After, Control-Impact; Underwood 1991, '94).  This method provides for a robust evaluation of environmental differences over time between an upstream and downstream site (Green & Montagna 1996) and tests for interaction between time and location (sampling sites) above and below an area of possible impact.  A significant interaction would indicate that both sites are not changing similarly over time, suggesting an effect of the silviculture/BMP activity on the stream ecosystem.  Multiple control sites allow for spatial replication and averages out spatial variation, thereby increasing the confidence that any interaction that is detected is not due to differences in the pattern of temporal variation between sites, but rather due to an effect of the silviculture activities.  In other words, multiple control sites will reduce the chance of a false positive (finding a significant effect of the silviculture activity when in fact no effect is occurring).  Multiple downstream sites would not be useful (except as backup) because such sites are expected to differ since any impact on a stream would decrease as the distance downstream from the impact site increased. 

For each of the eight silvicultural sites, sampling sites (pool/riffle combination) will be established immediately upstream of the proposed cut and immediately downstream of the proposed cut, for a total of 32 sampling sites.  Additional downstream sites may be sampled but not analyzed to serve as backups for unforeseen natural or anthropogenic events.   These sites will each be sampled twice per year in late spring/early summer and mid-winter.  This will start the summer before harvesting of timber begins within each silvicultural site. Samples will again be collected at each sampling location immediately before and immediately after the harvest has been completed and continue twice a year for two years.  Multiple sampling during the year will account for temporal variation among sites.  This sampling protocol will test for both short-term and long-term effectiveness of silvicultural BMPs and provide for statistically defensible data.

 During each sampling period, three replicate stations within each riffle and pool will be sampled using dip nets, with downstream sampling always occurring first.  Benthic macroinvertebrates will be sampled from multiple substrates (snags, leaf packs, root mats, bottom sediment, etc.) using 20 discrete dip net sweeps that will sample all stream habitats in the proportion in which these habitats occur at an individual site.

 All benthic macroinvertebrates (including chironimidae) will be identified to at least the genus level (specific level if a key exists for a particular species) using appropriate microscopy techniques by a qualified taxonomist (Ph.D. - candidate graduate student recruited from the North American Benthological Society or professional contacts).  These data will be used to determine multiple indices of  stream biological condition, including taxa richness, percent dominance, numbers of pollution sensitive taxa, EPT taxa, community structure and trophic distribution/feeding guilds and stream condition index (SCI).  SCI is a composite water quality and bioassessment index that is often useful for determining baseline levels and changes in stream ecosystem health.  The SCI works by assigning quality points to specific variables based on a comparison with an expected reference condition.  The Florida study used an SCI developed specifically for that state (Barbour et al 1996).  The present proposed study will utilize the SCI developed in Arkansas by the AR Department of Environmental Quality, which has been shown to be sensitive to changes in stream habitat characteristics in the Ouachita and Ozark ecoregions of Arkansas (C. Davidson (ARDEQ), personal communication). 

 Physical and chemical variables will be assessed at each sampling station, including flow, turbidity, particle size distribution, pH, dissolved oxygen, conductivity, temperature and nutrients (nitrate, phosphorus).  Nutrients will be measured spectrophotometrically using a commercially available field analytical system (HACH company).

 Habitat quality at each sampling station will be assessed and scored at each sampling station.  This will include bottom substrate/available cover, water velocity, artificial channelization, habitat smothering, bank stability, riparian zone buffer width and riparian zone vegetation quality.

 The data will be stored in electronic spreadsheets and statistically compared using Minitab statistical software.  Differences over time (before and after silviculture and BMP implementation) in measured physical, chemical and biological variables between upstream and downstream sites will be tested using two-way analysis of variance (after suitable data transformation, if necessary) using the following model

 X_ijk = m + a_i + t_k(_i) + b_j  + (a b)_ij   + e_ijk

 where m  is the overall mean,

a_i is the effect of period (_i = before or after silviculture activity), 

t_k(_i) represents sampling times within each period (winter or spring),

b_j  is the effect of location (above or below silviculture site),

(a b)_ij is the interaction between period (before or after) and location (above or below), e_ijk  represents the remaining error (variation)  for each data point X_ijk

The use of the parametric analysis of variance is greatly preferred over a non-parametric approach (e.g. Friedman's test) for this project because of the relatively small number of samples that will be collected.  For small sample sizes, parametric tests are much more powerful (able to detect a true effect) even when the data distribution is not normal (Boneau 1962).

Depending on the statistical expertise of the graduate student,  multivariate statistical analyses (i.e. principal components analysis, cluster analysis) may be employed as appropriate.

Reports will be made on a quarterly, annual, and final basis.  The Arkansas Forestry Commission and Arkansas Soil & Water Conservation Commission will use these reports to measure ongoing success of the project and to adjust as necessary to ensure continued success in meeting the goals of the project.

Literature cited

Adams, T.O., D.D. Hook and M.A. Floyd.  Effectiveness monitoring of silviculture best management practices in South Carolina. J. Appl. For. 19(4):170-176.

Barbour, M.T., J. Gerritsen, and J.S. White. 1996.  Development of the Stream Condition Index for Florida.  Prepared for the Fla. Dept. Environ. Protection. 105p.

Boneau, C.A. 1962. A comparison of the power of the U and t-tests. Psychological Review, 69:246-256.

Brown, A.V. and M.M. Lyttle. 1992. Impact of gravel mining on Ozark stream ecosystems. Final Report. AR Game and Fish Commission. Fisheries Division, Little Rock, AR.

Carline, R.F. and D. Lieb. 1996. Reduction of  agricultural nonpoint source pollution in the Spring Creek watershed, Centre County, PA. Final report. Centre County Conservation District, Bellefonte, PA. 27p.

Cobb, D.G., T.D. Galloway, and J.F. Flannagan. 1992. Effects of discharge and substrate stability on density and species composition of stream insects. Canadian Journal of Fisheries and Aquatic Sciences 49:1788-1795

FDEP. 1997.  Biological assessment of forestry best management practices. Florida Department of Environmental Protection. Bureau of Laboratories, Division of Administrative and Technical Services. 48pp.

Lynch, J.A. and E.S. Corbett.  1990.  Evaluation of best management practices for controlling nonpoint pollution from silvicultural operations.  Water Resour. Bullet. 26(1):41-52.

Mauney, M. and G.L. Harp. 1979. The effects of channelization on fish populations of the Cache River and Bayou DeView.  Proc. Ark. Acad. Sci. 33:54-54.

McClurkin, D.C., P.D. Duffy, S.J. Ursic, and N.S. Nelson. 1985.  Water quality effects of clearcutting upper coastal plain loblolly pine plantations.  J. Environ. Qual. 14(3):329-332.

McMahon, G. and T.F. Cuffney. 2000.  Quantifying urban intensity in drainage basins for assessing stream ecological conditions. J. Amer. Water Resour. Assoc. 36:1247-1255.

Johnson, R.L., D. Ward and R.S. Grippo. 2000. Physicochemical characteristics and macroinvertebrate assemblages of riffles upstream and downstream of a stream bank impacted by unrestricted cattle access.  AR. Acad. Sci. 54:68-76.

LaGasse, P.F., B.R. Winley and D.B. Simmons.  1980. Impacts of gravel mining on river system stability.  J. Waterways, Port, Coastatl and Ocean Div., ASCE 106:398-404.

Underwood, A.J. 1997. Experiments in ecology: their logical design and interpretation using analysis of variance. Cambridge Univ. Press.

Ursic, S.J. 1979.  Forestry practices and the water resource of the upper coastal plain.  School of Forest Resources and Conservation, University of Florida Resource Report No. 6.93-96. Gainesville, FL.

Green, R.H., and P. A. Montagna. 1996. Implications for monitoring: study designs and interpretation of results. Can. J. Fish. Aquat. Sci. 53:2629-2636.

 Underwood, A.J. 1991. Beyond BACI: experimental designs for detecting human environmental impacts on temporal variations in natural populations. Australian Journal of Marine and Freshwater Research 42:569-587.

Underwood, A. J. 1994. On beyond BACI: sampling designs that might reliably detect environmental disturbances. Ecol. Appl. 4: 3-15.

Ursic, S.J. 1991. Hydrologic effects of clearcutting and stripcutting loblolly pine in the coastal plain.  Water Resour. Bullet. 27(6):925-937.

Wynn, T.M. S. Mostaghimi, J.W. Frazee, P.W. McClellan, R.M. Shaffer, W.M. Aust. 2000. Effects of forest harvesting best management practices on surface water quality in the Virginia coastal plain.