15
II. DISSECTED ARCHEOLOGICAL SITE CLEARING
BACKGROUND Since the establishment of the Park in 1978, many resources have been dedicated to
vegetation management for protecting archeological features. These efforts continue today and are most reliant on physical removal actions, which are labor intensive. To better understand how labor resources are applied in these vegetation removal activities, we developed a Dissected Archeological Site Clearing (DASC) protocol to independently record each aspect of an operation from cutting to hauling. This protocol was then implemented at six different archeological sites towards development of a model for estimating labor resources for future site clearings.
METHODS Six archeological sites were cleared from 2009‐2012 mostly by a two‐person crew (on
one occasion D13‐34 clearing was achieved with a large volunteer crew) using DASC protocols (Table 2). Prior to recording any vegetation removal activities, the boundaries of each site encompassing the archeological features were delimited and mapped, along with establishment of semi‐permanent photo points. Features were made visible with flagging to ensure that none were disrupted during the removal process. No less than four photo points were positioned around the perimeter of the site and one panoramic photo was taken from the center of the site. The site was delimited by a perimeter that was established no less than five meters from any one of the designated features. In certain circumstances, the perimeter may have been extended to accommodate removal of other nearby vegetation encroaching upon the site.
16
Figure 6. Photos of a clearing‐site (A) before clearing (B) immediately after clearing (C) 6 months after clearing. Photos were taken from the same photo point facing the same direction.
B
A
C
17
Table 2. The six archeological sites cleared during the study.
Site Area (m2) Start Date End Date
D13‐10 1,034 Jan‐2009 Jul‐2009
D13‐58 2,917 Sep‐2009 Jul‐2010
D13‐40 244 Dec‐2010 Mar‐2011
D13‐39 234 Dec‐2010 Apr‐2011
D13‐32 756 Apr‐2011 Jul‐2011
D13‐34 “C” 999 Jan‐2012 Jun‐2012
Time was recorded with a stopwatch for each activity from start to completion with stops in between when the activity was halted (i.e. break). Net cumulative time was reported for each activity. All minutes and hours reported herein are “person‐minutes” and person‐hours”. Cutting and hauling procedures were determined by the vegetation type and size. Typically trees and shrubs were cut using a chainsaw and hauled from the site using a utility task vehicle (UTV) while grasses were cut with a line‐trimmer, but was not removed from the site. Some sites required additional activities such as the use of pole‐saws in order to limb larger trees before felling. Vegetation cutting commenced by order of vegetation type and size: (i) small‐class kiawe (<20 cm diameter at breast height (DBH)), (ii) small‐class Christmas berry (< 20 cm DBH), (iii) large‐class Christmas berry (>20 cm DBH), (iv) grass suppression, and finally (v) large‐class kiawe (>20 cm DBH), if necessary. This order was maintained in our trials for comparative purposes. Other DASC protocols included cutting stumps to within 5 cm of the ground, cutting branches to approximately 50 cm for consistent ease of hauling, and piling cut vegetation no more than 50 cm high for the same reason. Cut tree stumps received an herbicide application (triclopyr at 1‐2 lbs. active ingredient per gallon). Cut material were stacked and readied for hauling. If piles were located near UTV loading sites, then they were free loaded. If piles were greater than 5 m from the UTV, they were delivered to the UTV in a 0.12 m3 receptacle (32‐gallon trash can). Vegetation could often be stacked above the height of the cargo bed, so visual estimates of the total number of cargo beds full of vegetation were made for each UTV load. For example, if vegetation extended twice the height of the cargo bed, then the payload was equal to two UTV beds of vegetation. UTV payloads were recorded each trip because they contained varying combinations of logs and brush; however, we estimated an average payload to be 0.63 m3. Grasses were also cut within 5 cm of the ground and spot sprayed with herbicide (glyphosate or imazapyr at 0.04‐0.08 lbs. ai/gal; 2% v/v). This step was usually the last activity that took place 2‐3 weeks after cutting and hauling the woody components from the site.
The thick vegetation cover in site D13‐34 made it too difficult to perform a pre‐clearing
survey or to remove vegetation components in the order of the DASC protocol. Instead, inventory and vegetation measurements were collected concurrently with removal activities. A grid was established within the D13‐34 clearing site to keep track of the vegetation inventory and to help estimate removal time. A total of 31 sections (5 x 5‐m; ca. 775 m3) divided up the stand. Prior to removing a section, average height and a visual estimation of percent cover
18
were recorded. Counts of individual plants per grid‐square were recorded if plant occurrence and growth form allowed. In addition, all basal diameters of vegetation were recorded, excluding basal diameters <10 cm for Christmas berry and kiawe. Time requirements for cutting and hauling of vegetation at D13‐34 were recorded separately within each grid square. Chemical treatment of cut stumps was also included in cutting time.
RESULTS D13‐34
The vegetation covering Archeological site D13‐34 area “C” consisted almost exclusively of Christmas berry and grew in dense, horizontal vine‐like branches 2‐4 m in height from the ground (e.g., see Figure 7).
The percent cover of Christmas berry over the area, before clearing, was 84% with an average height of 2.5 m (range 1‐4 m). Besides the other exotic species, there were also some native species, and Polynesian introductions represented: Thespesia populnea (milo), Scaevola taccada (naupaka), Morinda citrifolia (noni), and Capparis sandwichiana (maiapilo). None of these contributed >1% cover (Table 3).
Mean cutting time among the 31 sections was 3.6 ± 1.6 min/m2 (mean ± SD), with
variability mostly along the edges of the infestation where it inherently thins out. Excluding these edge sections, mean cutting time increased to 4.2 ± 1.3 min/m2. Christmas berry removal was 93% of the total cutting time.
Each 5x5 m grid square contained an average of 2.23 ± 1.04 m3 of vegetation translating
into an average of 3.4 UTV loads. At distances between 5‐30 m, loading time (i.e., time required to move the vegetation from the cut pile to the UTV bed) was positively correlated with increasing distance (R2= 0.54) between the UTV and the pile (data not shown). On average, loading one UTV took 34.7 person‐minutes with 1.0 additional minute for every meter of distance between the pile and the UTV. This trend was observed using the receptacle for hauling brush and free loading logs. When distance between the piles and UTV were short enough to free‐load brush into the bed of the UTV (i.e. <5 m), loading time was greatly reduced, averaging 16.6 ± 7.3 person‐minutes per load.
UTV travel‐distance to the disposal site was 986 meters and took 10.3 ± 2.5 person‐minutes round trip. A total of 110 UTV loads hauled away an estimated 73.4 m3 of vegetation to the disposal site. Unloading vegetation at the disposal site took an average of 6.3 ± 2.0 person‐minutes per load, for an added 11.6 person hours. For this site, UTVs traveled a total of 217 kilometers in 37.7 person‐hours. Total loading/hauling time was 118 hours; 71% of the total time (Table 4). Travel between the Park maintenance yard and the clearing site (not included in
19
Table 3. Plant species found within clearing boundary of site D13‐34.
Species StatusMax Ht (cm)
# of individuals
Sum Dia. (cm)
% coverage
Christmas berry A 400 74* 1868.5* 84% Kiawe A 240 2* 16.75* 2% Bittermelon M. charantia A NA 2 NA <1% Maiapilo C. sandwichiana E 225 2 17.7 <1% Naupaka S. taccada I 120 NA NA 1% Noni M. citrifolia P 470 14 67.6 <1% Milo T. populnea P 700 3 32.1 3% Fountain grass A 120 NA NA 10‐25%**
Notes: Status A = Alien, E = Endemic, I = Indigenous, P = Polynesian introduced: * includes only basal diameters >10 cm ** approximately 100m2 of 10‐25% grass cover on eastern side of the grid. DASC timetable) added another 66 ± 10 person‐minutes of UTV operation per day when using the Mamalahoa Trail and Hu’ehu’e Ranch Road. The Ala Kahakai Coastal Trail was faster, averaging 40 ± 5 person‐minutes roundtrip. Site D13‐10
Clearing activities for site D13‐10 began in January 2009. The northern portion of the site was cleared in a previous activity months earlier, so the southern portion was the focus of this particular clearing activity. A rock wall feature outlined the cultural site. The total cleared area, including a 5‐m buffer around the rock wall, was 1034 m2. UTV access was within 5 meters, allowing for vegetation to be free loaded.
Small and large class Christmas berry was by far the dominant vegetation and access to
the dense thicket impeded progress (see Figure 7). In total, 48 person‐hours were required to remove this species, which accounted for 72% of all vegetation clearing and 40% of the total time (Table 5). Loading/hauling of the vegetation required 55 person‐hours (45% of total). In total, this site was cleared in 14 work days within a six‐month period that ended in July 2009.
20
Table 4. DASC timetable and order of clearing for site D13‐34. Fountain grass not included in loading/hauling.
DASC Order Person‐ Minutes
Person‐ Hours
Person‐Work Days
Christmas berry 2761 46 6 Kiawe 61 1 0.1 Fountain grass 156 3 0.3 Loading/Hauling 7062 118 15
TOTAL 10,040 168 21.4
Table 5. DASC timetable and order of clearing for site D13‐10.
DASC Order Person‐minutes
Person‐Hours
Person‐Work Days
Tagging trees 66 1 0.1 Small kiawe 455 8 1 Small Christmas berry 1371 23 3 haole koa 122 2 0.3 Large Christmas berry 1488 25 3 Large kiawe 172 3 0.4 Other 218 4 1 Loading/Hauling 3298 55 5
TOTAL 7190 121 14
21
Figure 7. Growth form of dominant Christmas berry observed impacting many of the sites
Site D13‐58
Clearing activities for D13‐58 began on September 2009. This particular site was the largest, encompassing 2917 m2, with 14 individual archeological features identified. Due to its size, and issues with time, we couldn’t clear cut one large swath; instead we had to clear pockets around the features. Kiawe was the dominant species at this site, taking 55 hours to clear, which was equivalent to 45% of the vegetation clearing and 16% of the total time (Table 6). Loading/hauling of the vegetation required 218 person‐hours (64% of total). This site was completed in 42 work days within a 10‐month time period that ended July 2010.
Table 6. DASC timetable and order of clearing for site D13‐58.
DASC Order Person‐minutes
Person‐Hours
Person‐Work Days
Small Christmas berry 753 13 2 Small kiawe 876 15 2 Large Christmas berry 50 1 0.1 Haole koa 985 16 2 Large kiawe 2423 40 5 Fountain grass 1607 27 3 Other 577 10 1 Loading/Hauling 13085 218 27
TOTAL 20,356 340 42
22
Site D13‐32 Clearing activities commenced in July, 2011 and were completed 4 months later in August. Clearing activities mainly involved felling and hauling a large kiawe tree, which was 97% of the 12 work‐days required to complete the task (Table 7). Site D13‐39 Site D13‐39 was the smallest of the clearing sites at 234 m2. This site was another example of a removal operation of a single large‐class species (Christmas berry) that was completed in 6 work days. It should be noted that 79% of the time was dedicated to hauling the materials away from the site, which is due to long distances to the UTV and the disposal site (Table 8). Site D13‐40 Site D13‐40 was the second smallest site and again dominated by large‐class specimens, in this case Christmas berry and kiawe. This was also another situation where hauling time was a substantial amount of the effort at 82% of the total time (Table 9).
Table 7. DASC timetable and order of clearing for site D13‐32.
DASC Order Person‐minutes
Person‐Hours
Person‐Work Days
Large kiawe 2364 39 5 Fountain grass 101 2 0.3 Loading/Hauling 3231 54 7
TOTAL 5,696 95 12
Table 8. DASC timetable and order of clearing for site D13‐39.
DASC Order Person‐minutes
Person‐Hours
Person‐Work Day
Large Christmas berry 479 8 1 Fountain grass 72 1 0.1 Other 26 0.4 0.05 Loading/Hauling 2204 37 5
TOTAL 2,781 46 6
23
Table 9. DASC timetable and order of clearing for site D13‐40.
DASC Order Person‐minutes
Person‐Hours
Person‐Work Day
Large Christmas berry 458 8 1 Large kiawe 202 3 0.4 Loading/Hauling 2941 49 6
TOTAL 3,601 60 7
DISCUSSION Leaving the vegetation at the site can continue to be an obstruction to archeological
features and also cause a fire hazard, making this practice non‐compliant with Park policy. With that in mind, removal of vegetation from the sites is a substantial financial obligation. In our assessments, loading and hauling activities accounted for 46‐81% of the total time among the DASC sites, suggesting that most of the improvements in efficiency may result from changes in these activities. The capability to park the UTV within 5 m of the clearing site allowed for free‐loading vegetation was a major time saver. However, to protect cultural and natural resources, UTVs must remain on access trails. It should be noted that timetables for each clearing site are only specific to the physical acts of cutting and hauling vegetation. Secondary activities (e.g. equipment maintenance, repair, hydration breaks, travel between maintenance yard and site, travel between park and office, cleaning vehicles, etc.) were not included in the timetables, but we estimated that this could take up to an additional four hours each day. For instance, when a site is located on the north side of the Park, travel from the Park maintenance yard to the clearing site each day consumed over one hour per person, One solution may be to consider satellite areas for temporarily securely storing equipment for more efficient access to some of the remote locations on the Park. For site D13‐34, being able to temporarily store the UTV and cutting equipment near Kaloko Pond could have saved up to 60 person‐hours (i.e., 36% of total effort). Thus, labor estimates should double when estimating total workdays needed to clear a site. Recommendations: All of these sites were dominated by large‐class specimens that required the majority of labor resources for removal. These sites will require follow‐up vegetation control activities. Although follow‐up control was not accounted for in this study, we anticipate that rescheduling maintenance when the vegetation is smaller will be a more efficient use of labor resources. In the following section, DASC time records from these were analyzed and a model was constructed to predict time requirements for future removal efforts pertaining to large‐class dominated vegetation.