The soil map is based on 484 dedicated soil survey auger observations at 443 sites and 30 detailed soil profile descriptions. The augering for the soil survey was done in three stages – main survey, boundary checking, and spatial variability testing (Table 4.1).
There are a further 273 sites at which the soil classes were identified by members of the soil survey team in the course of other studies (e.g. Barthold, F., R.F. Stallard, and H. Elsenbeer. 2007. Soil nutrients and landscape relationships in a lowland tropical rainforest in Panama. Submitted to Forest Ecology &. Management.)
Soils classes have been identified in > 750 observations at > 700 sites (Figure 4.1), giving a density of slightly less than 0.5 per hectare. This qualifies the survey as 2nd order (Soil Survey Staff, 1997; Schoenberger et al. 1998), level 2 detailed (Dent & Young, 1981), detailed (FAO, 1979) or semi-detailed (Landon, 1991).
Table – Summary of soil auger observations, 2005-6
|Date of fieldwork||Designation||Purpose||Basis of location||Number||Data collected|
|March – April 2005||Main survey||Establishment of soil classes & soil mapping units, & drafting soil boundaries||Subjective free survey; mainly short topographic transects from trail markers||
|Site & horizon description to 1.5 – 2 m|
|April 2005– March 2006||Boundary check||Refinement of soil boundaries||Subjective, in areas of pedocartographic uncertainty||
|Site & horizon description to 1 m, and soil class ID|
|March 2006||Test||Validation and quantification of the variability of mapping units||7 sites in major mapping units; with 8 nested dumbbell random walk augerings at 1, 10 & 100 m intervals at each site.||
|58||Site & horizon description to 1 m and soil class ID|
|Additional auger data from non-soil survey projects|
|May – September 2005||Soil nutrients
SOC & clay mineral studies
|Characterisation of soils sampled||Stratified random||
|Soil class identified|
* Some sites have multiple augerings where shallow stones stopped first attempts
The soils were examined in the main survey on a free survey basis, with sites subjectively chosen where forest and landscape indicate soil differences (Jenny 1980; Hudson, 1992). We used the trail network and 100 m markers for access and navigation, and most augerings were near trails. The trails tend to follow ridge and spur crests, especially in the steeper terrain in the northern part of the island, although Svenning et al. (2004) found that the topographic bias of the trail system is less than first appears. Nonetheless, we consciously avoided the tendency to over-sample ridges and upper slopes by augering along short topographic transects, usually at three points – crest, midslope and drainage line/lakeshore – orthogonal to the trails at the subjectively chosen points. The additional check augerings in 2005 were sited subjectively to resolve particular problems of soil boundary location on the draft map.
The density of augerings was higher in areas of particular ecological and hydrological research interest. Such as the 50 ha LTER plot, and the catchment of Lutz Creek and the upper catchment of Conrad Creek. This means that the soils of the andesite plateau and the northern outcrop of the Caimito marine facies are sampled at higher than average densities.
Morphological variation within soil mapping units test was characterised in 2006 at seven sites, at each of which eight soil augerings were positioned at 1, 10 and 100 m intervals in nested, random-bearing, dumbbell patterns (Figure 4.2). These observations were intended to determine the proportions of main and minor soil classes in the soil-mapping units. (Forbes et al. 1984), and the spatial scale of soil morphological variation.
All augerings were done with 7 cm diameter Terrasystem (USA) or Edelman (Eijkelkamp, Netherlands) augers. Because the BCI dry season imposes moisture stress on the vegetation, particular attention was given to the depth of the lithic/paralithic contact and nature of the sub-solum material. Some soils were augered to 2m in the main survey, and the rest to 1.5m, unless stopped by unequivocal saprolite or hard stones. Where augering was stopped at less than 50 cm by hard stones, up to two further attempts were made nearby (at < 3m).
Auger descriptions included matrix colour, mottles, hand texture, stones, concretions, moisture content and auger consistence for all natural horizons. Hand texturing tends to overestimate silt contents in humid tropical fine textured soils (Akamigbo, 1984). However, the high silt estimates in our field textures are corroborated by laboratory analyses, even with intensive dithionite dispersion. Our field estimates of moderate – high clay contents are corroborated by laboratory data for both sedimentation and water retention (Soil Survey Staff, 1999).
We more fully characterised the main soil classes (FAO, 1990; Soil Survey Staff 1993) in 30 subjectively sited soil profile pits. These were dug to depths of 1.5 – 1.7m, with augering to a further 2m in the bases, giving total observation depths > 3.5m where necessary. The pit faces were sampled in the topsoil and then by 10 cm depth increments. The augerings in the bases were not sampled. The samples were oven-dried at 60º C on BCI until touch-dry. The dried samples were subdivided, with half of each going to the BCI archive, and half to Germany.
Linked interactive databases with the full profile, auger and analytical data will be posted on the STRI website. They will be freely available, on the understanding that the data were collected at subjectively chosen sites for a free soil survey. People wishing to do statistical analyses of soil attributes for other studies will often need to collect their own data.