United States Department of Agriculture
Forest Service
Pacific Southwest Research Station

General Technical Report

Persistence of Historical Logging Impacts on Channel Form in Mainstem North Fork Caspar Creek1

Michael B. Napolitano2

1An abbreviated version of this paper was presented at the Conference on Coastal Watersheds: The Caspar Creek Story, May 6, 1998, Ukiah, California

2Environmental Scientist, San Francisco Bay Regional Water Quality Control Board, 1515 Clay Street, 14th Floor, Oakland, CA 94612 (mbn@rb2.swrcb.ca.gov)

Abstract: The old-growth redwood forest of North Fork Caspar Creek was clear-cut logged between 1860 and 1904. Transportation of logs involved construction of a splash dam in the headwaters of North Fork Caspar Creek. Water stored behind the dam was released during large storms to enable log drives. Before log drives could be conducted, the stream channel had to be prepared by removing all obstructions, including large woody debris jams, from the channel. Comparison of present-day woody debris loading on North Fork Caspar Creek (24 kg m-2) to physically similar streams in old-growth redwood basins (49 to 268 kg m-2) suggests that wood-loading and stability were greatly diminished by historical logging activities and change to second-growth cover. These changes are important, as woody debris creates large-volume, long-term sediment storage sites and diverse aquatic habitat conditions. Although historical logging appears to have caused lasting channel changes, including channel incision, simplification of form, and reduction in sediment storage capability, the significance of habitat-related changes remains unclear.

eller and Tally (1979), in a study of the role of woody debris in steep, headwaters streams draining old-growth redwood forests, found that debris provides: (a) a stepped channel profile where a large proportion of the stream's total energy is dissipated locally at plunge pools below debris dams; (b) stable channel roughness elements that provide large-volume, long-term sediment storage sites (often stable for hundreds of years), effectively buffering the channel from infrequent large sediment inflows; and (c) stable channel structure that creates a diverse assemblage of channel morphologies and flow conditions.

  Stable and diverse channel form is often associated with high-quality fish habitat. Physical factors (stream order, discharge, valley width, channel type, channel slope), woody debris input processes, and the size of debris pieces interact to control frequency, distribution, and stability of in-stream woody debris over time (Keller and others 1981). The influence of woody debris on channel form and process is directly related to its amount per unit length (debris loading), distribution, and stability over time.

  To evaluate whether historical logging has caused persistent changes in channel form, woody debris loading, and stability, I analyzed: (a) research regarding the effect of woody debris on channel form and function in streams draining second- and old-growth redwood forest; (b) history of 19th- century logging activities at Caspar Creek; and (c) field evidence for historical disturbance or removal of wood from North Fork Caspar Creek. This paper describes analysis of these data and discusses probable channel response to 19th- century logging activities.

Site Description

The mainstem channel of North Fork Caspar Creek, located in Mendocino County, California, is a steep (slope = 0.02), perennial, gravel-bed stream that is confined within a deeply-incised inner gorge. Position of bank-side trees and occurrence of large woody debris strongly influence channel position, variability in form, and width. Most sediment within the active channel is stored: (a) as localized deposits associated with jams of large woody debris; and (b) along short reaches of channel that are aggrading and widening in response to adjacent recent landslides. Gravel bars in the mainstem channel are unvegetated or covered with short-lived hydrophytes. Valley fill terraces define one or both channel banks along most of the channel length and become increasingly common downstream. Old-growth stumps in growth position on many valley fills confirm that some terraces were deposited at least hundreds of years ago, and that bank erosion and channel migration rates have subsequently been very low.

Comparison of North Fork Caspar Creek to Similar Streams in Old-Growth Coast Redwood Forest

Research by Tally (1980) demonstrates that much of the variability in debris loading along a particular stream draining an old-growth redwood forest is related to frequency of "large diameter redwood trees" (table 1) that are located near the channel. When physical input factors are uniform, debris loading is primarily a function of tree frequency, and therefore, physically similar channels should have comparable debris loading given similar forest cover.

  Before 19th-century logging, tree frequency on North Fork Caspar Creek is likely to have been within the range for steep mountain streams (e.g., those without extensive floodplains) in old-growth forests that were surveyed by Tally (1980). Tree frequency along these streams varies from 26 to 68 trees per hectare. Keller and others (1981) compared several streams in second- and old-growth redwood basins to assess how the influence of woody debris on channel form and process may be altered in second-growth basins (table 2). North Fork Caspar Creek was one of the second-growth basins studied. Keller and others (1981) estimated debris loading of 21 to 24 kg m-2 in North Fork Caspar Creek. Of the old-growth streams studied by Keller, upper Little Lost Man Creek is the most similar to North Fork Caspar Creek (table 3). Both are steep, second-order, gravel-bedded streams with narrow valleys, similar drainage area, channel width, and slope. Therefore, physical factors affecting woody debris input and loading should be similar. There are between 26 and 52 large redwood trees per hectare along Little Lost Man Creek; the creek contains 49 to 268 kilograms of wood per square meter of stream (table 4) or two to seven times more than in North Fork Caspar Creek. Therefore, it appears that debris loading, and consequently the influence of woody debris on channel form and process, was much greater in North Fork Caspar Creek before the 19th- century logging, compared to present conditions. Much higher debris loading in Little Lost Man Creek provides significantly greater debris-related sediment storage capacity (table 4). For example, log jams in Little Lost Man Creek store about five times as much sediment, and have approximately 20 times as much unfilled storage capacity as in North Fork Caspar Creek.

Table 1-Large woody debris loading in streams draining old-growth redwood forests. (Source: Tally 1980).
(kg m-2)
Number of
large redwoods
within 50 m
Hayes Creek17068None
Little Lost ManUpper14152None
Prairie CreekHope Creek21880Minor
Little Creek1225Yes
Forked Creek1321Yes
Zig Zag No. 22225Yes
Natural Tunnel10641Minor
Brown Creek8575None
r2 = 0.88 for debris loading vs. large redwood frequency.

Table 2-Channel attributes for streams in second-and old growth redwood forests. Source: Keller and others 1981).
North Fork
Caspar Creek
Little Lost
Man Creek
Zig Zag
No. 2

Basin area (km2)1.6/
Stream order2/22322/22222234
Debris loading
(kg m-2)
Pool to pool spacing
(by channel widths)
Area in pools (pct)24/3633271222/1849344636412625
Area in riffles (pct)30/3025142615/2121464920151825
Area in debris-stored
sediment (pct)
Area in undercut
banks (pct)
2/14243/114341< 11
Pool morphology
influenced by
debris (pct)
Debris controlled
drop in elevation
57/3769173859/304327348< 118< 1

NOTE: Total percentages in stream environments may be less than or greater than 100 percent owing to overlaps between units (such as pools which contain debris-stored sediment).

  Woody debris jams in streams that drain old-growth forests are also quite stable. Large trees, found growing on pieces of debris which comprise the jams, are often more than 100 years old (Keller and Tally 1979). Considering the stability of debris jams in old-growth streams, Keller and Tally (1979) concluded that debris-related sediment storage capacity in Little Lost Man Creek (and other old-growth streams) provides an important buffer system for the channel by allowing infrequent large-magnitude sediment inputs to be stored in jams and released slowly over time. In contrast, debris-related storage capacity in North Fork Caspar Creek is less than 50 t km-2 (table 4), the presence of many collapsed or partially collapsed jams and the lack of mature trees growing through the debris pieces suggest that the debris jams are dynamic, short-lived features (Napolitano 1996; table 5). Historical logging activities may be the cause for these differences.

History of 19th-Century Logging at Caspar Creek

Caspar Creek was first logged in 1860, and most of the watershed was clearcut and burned between 1864 and the mid-1890's (Wurm 1986). Caspar Lumber Company records indicate that redwoods logged in the Caspar Creek watershed typically ranged between 0.8 and 2.5 m in diameter. Cut logs were floated downstream to the company mill located on the coast. To make this possible, a logging splash dam was constructed near the headwaters of the North Fork Caspar Creek (Jackson 1987a). The water stored behind the dam was released during large storms to increase streamflow enough to enable log drives. Before log drives could be conducted, a stream channel had to be "improved" by "removal or blasting of boulders, large rocks, leaning trees, sunken logs or obstructions of any kind" (Brown 1936). During each log drive thousands of logs were transported down the creek (Jackson 1987b).

Table 3-Comparison of characteristics of Upper Little Lost Man Creek and North Fork Caspar Creek1.
growth slopes or
growth valley
flat and/or

1Data for Little Lost Man Creek from Keller and Talley(1979)
2Mean channel width=channel area per centerline channel length

Table 4-Comparisons of characteristics of Upper Little Lost Man Creek and North Fork Casper Creek1
Woody Debris
Upper Little
Lost Man
North Fork

1 All Little Lost Man Creek data, and debris loading data for North Caspar Creek are from Keller and others (1981).
2 Remaining sediment storage capacity in debris jams.
3 Based on data in Keller and others (1981), and assuming sediment storage per unit drainage area is similar in upper and lower Little Lost Man Creek and bulk density of sediment in storage is approximately 1.8 t m-3.
4 North Fork Caspar Creek sediment storage based on data collected in summer 1987.

Field Evidence of Channel Improvement and Log Drives

Evidence of channel preparation for log drives along the mainstem North Fork Caspar Creek can be found by examining in-place old-growth stumps on valley fills. The old-growth redwood stumps are commonly obscured by mature stump sprouts or by shrubs growing through the stump. It is likely, therefore, that old-growth stumps are present elsewhere along the creek where they have not been recognized. As the valley width is narrow (3 to 20 m) along most of North Fork Caspar Creek, stumps were cut flush with the ground surface to avoid snagging of floated logs during drives. All other old-growth stumps in the basin (e.g., those farther from the channel and on hillslopes) were cut well above the root swell, several meters above ground surface, because sawyers were paid by the small diameter of each log that they cut (Jackson 1987a).

  Direct evidence of removal of woody debris elements from the channel of North Fork Caspar Creek is difficult to find. Characteristics of woody debris within the active channel, however, suggest that logs were removed or blasted. For example, almost without exception, the largest logs in the channel today are 0.5 m in diameter, approximately the same diameter, as the largest second-growth trees within the basin. In one location, an old-growth trunk is protruding from the bank of a valley fill deposit. This trunk had been sawed obliquely, to be flush with the ground surface of the valley fill deposit. Before being cut, it probably extended across the valley width and obstructed streamflow, and thus would have hindered efforts to float logs downstream. Other smaller old-growth logs are similarly oriented and partially buried within the same valley fill deposit a few meters upstream, suggesting that there may have originally been a debris jam present at the site.

Table 5-Large debris jams in North Fork Caspar Creek having sediment storage volume 25m3.
map I.D.
LocationDebris jam
water year
in storage1
Evidence from maps1
0180 m upstream
of xs 9
1980530-10m3Jam formed in 1980, as noted in 1980 xs survey;
bars and some LD first depicted on 1986 map
0225 m upstream
of xs25
1984 or 19853420-30 m3increaseLWD jam but no bars on 1985 map; xs 26 end-pins
missing in 1986; step and small bar shown on 1986 map
0315 m downstream
of xs 28
Before 19792710-10 m3Few LWD pieces and no gravel bars on 1985 map;
long bar on 1986 map
058 m upstream
of xs 37
Before 19792580-10 m3Most bars and LWD are depicted on 1985 map;
no significant changes 1986-87
062 m downstream
of xs 42
1979 and 19853
730-10 m3Stepping noted 1985; step breached 1986, but most
stred sediment remained in jam
0715 m upstream
of xs 43
Before 1979247No changeNo changes evident 1985-87
01416 m upstream of
xs 50
1979 and 19853
77No changeNo changes evident 1985-87
01717 m downstream
of xs 56
Before 19792320-20 m3 decreaseStep collapsed in 1986,
but most stred sediment remained in jam
02426 m upstream
of xs 60
Before 1979226No changeNo changes evident 1985-87
03316 m downstream
of xs 74
Before 19792330-20 m3decreaseNo changes evident on maps; 1986-88 scour
at xs 74 suggests a decrease in storage
03517 m downstream
of xs 76
1979 and 19853
27No changeNo changes evident 1985-87
Total storage as Large Jams (530 m3)

1 Based on analysis large woody debris maps (Unpublished USDA Forest Service maps) prepared in 1985 and 1986, and geomorphic maps prepared in 1987 (Napolitano 1996).
2 Based on review of cross-section field notes prepared in July 1979, which state whether large woody debris was present, and if it created a backwater at a cross-section.
3 No backwater effect from woody debris noted in cross-section field notes prepared in July 1979; debris jam shown on 1985 large woody debris maps.

Channel Response to 19th-Century Logging Activities

Channel erosion and incision would be promoted by increased peak flows associated with splash dam releases and abrasion caused by repeated transport of thousands of logs. A large fraction of the sediment stored in debris-jam backwaters would probably have been liberated because the logs that had stabilized and trapped the sediment were removed during channel preparation. Considering that the diameters of trees logged in Caspar Creek generally ranged between 0.8 and 2.5 m, the streambed may have degraded substantially where jams extended across the channel. Most of the sediment stored in valley fills, however, probably was not eroded because of the resistance to erosion afforded by large and extensive root networks of the old-growth trees growing on the fills.

  Before the log drives, the mainstem channel is likely to have more closely resembled the present-day stream reach located upstream of the splash dam backwater. In that reach, the channel is only slightly entrenched (typically channel banks are less than 0.6 m high) and has a much higher width-to-depth ratio than below the splash dam. Its planform, typically, is anastomosing with a well-defined main channel and auxiliary high-flow channels.

  Under present conditions, the largest second-growth trunks in the channel in the reach upstream of the splash dam do not appear to be mobilized by frequently occurring peak flows. Interactions between the forest and the channel in that reach are more likely to resemble those before the initial logging than would the interactions downstream where the logs are more easily mobilized.

  Channel morphology in the reach above the splash-dam resembles that of Little Lost Man Creek, the old-growth channel in Redwood National Park which is similar to North Fork Caspar Creek in setting and physical watershed characteristics.

  Lack of well-developed soil horizons on the valley fills suggests that the fills were frequently flooded, at least as recently as several hundred years ago (i.e., the time it would take for a A horizon to form). The fact that old-growth trees on the valley fills were cut flush with the ground surface suggests that those preparing the channel for log drives believed this was necessary to avoid snagging cut logs during drives, also suggesting that high flows regularly inundated the terrace surface. Bank tops along North Fork Caspar Creek are typically 1 to 2 m above the channel thalweg, much greater than stages associated with common flows (i.e., a stage of about 0.6 m has a recurrence interval of 6 yr at gaging Station A). This suggests that valley fills have been converted from large-volume, long-term sediment sinks (floodplains) to substantial sediment sources (terraces) as a result of channel incision in response to removal of old-growth debris jams from the channel during 19th-century logging activities. Conversion of the floodplains to terraces signifies a major change of trends in valley sediment storage and a pervasive alteration in the sediment budget for the basin.

  The channel has not recovered its previous morphology because jams in the channel are now less stable, stepping is less pronounced with smaller diameter trunks, and the resistance to bank erosion afforded by second-growth trees on the valley fills limits lateral migration. These factors cause the channel to remain entrenched, and to have a narrower width-to-depth ratio than the reach above the splash dam. Comparison of second-growth to old-growth channels also shows that pools are much more frequent and their average depth is greater in the old-growth channels (Keller and others 1981, Montgomery and others 1995). Therefore, it is also likely that pools are less frequent and shallower in North Fork Caspar Creek as a result of historical logging activities. It is unlikely that North Fork Caspar Creek will recover its former morphology, however, until the former relationship between the size of woody debris and flow magnitude is reestablished.


Brown, N.C. 1936. Logging transportation. New York: John F. Wiley and Sons; 327 p.

Jackson, W. Francis, Historian and author, Mendocino, California. [In-person conversation with Michael B. Napolitano]. July 1987.

Jackson, W.Francis, Historian and author, Mendocino, California. [In-person conversation with Michael B. Napolitano]. August 1987.

Keller, E.A.; MacDonald, A.; Tally, T. 1981. Streams in the coastal redwood environment: the role of large organic debris. In: Coats, R.N., ed. Proceedings, Symposium on watershed rehabilitation in Redwood National Park and other Pacific coastal areas; 1981 August 25­28; Arcata, CA. Sacramento, CA: Center for Natural Resources Studies of JMI, Inc.; 161-176.

Keller, E.A.; Tally, T. 1979. Effects of large organic debris on channel form and fluvial processes in the coastal redwood environment. In: Rhodes, D.D.; Williams, G.P., eds. Adjustments in the fluvial system. Dubuque, IA: Kendell Hunt Publications; 168-198.

Montgomery, D.R.; Buffington, J.M.; Smith, R.D.; Schmidt, K.M.; Pess, G. 1995. Pool spacing in forest channels. Water Resources Research 31(4): 1097-1105.

Napolitano, M.B. 1996. Sediment transport and storage in North Fork Caspar Creek, Mendocino County, California: water years 1980-1988. Arcata, CA: Humboldt State University; 148 p. M.S. thesis.

Tally, T. 1980. The effects of geology and large organic debris on stream channel morphology and processes for streams flowing through old-growth redwood forests in northwestern California. Santa Barbara, CA: University of California at Santa Barbara; 273 p. PhD. dissertation.

Wurm, T. 1986. Mallets on the Mendocino coast: Caspar Lumber Company, railroads and steamships. Glendale, CA: Trans-Anglo Books.


Paper 1

Paper 2

Paper 3

Paper 4

Paper 5

Paper 6

Paper 7

Paper 8

Paper 9

Paper 10

Paper 12

Paper 13

Paper 14

Paper 15

Paper 16


Publishing Information