Skip to main content.

4. A National Monitoring System For Soil Carbon

4.1 System requirements

While forests may appear to be the most observable form of terrestrial carbon accumulation, soils are the major terrestrial carbon reservoir, the functioning of which can be disturbed in a variety of ways, predominately through land-use change and land-management practices.

As previously described, the monitoring system for soil carbon is based on the IPCC (1996) default methodology, default and follows the method used to update the 1990 baseline for soil C, and referred to in Figure 1. The country has been classified according to a matrix incorporating soil type reclassified to the IPCC guidelines, climate and land use.  Soil C is determined for three depth increments: 0–0.1 m, 0.1–0.3 m, and 0.3–1 m (Tate et al. 1997b).

The monitoring system is required to operate over a repeat period of about 10 years, in which time land use is likely to be the only major influence on soil C.  Hence, any future monitoring system should be designed to capture those changes in soil C brought about by a change in land use.  A system based on updating land use and applying land-change factors, known as coefficients of change (COC) is likely to be less costly than a plot-based sampling system providing there is a method to determine land-use change and providing the coefficients of change can be determined. If that were not to be the case, a plot-based sampling system, whether random or targeted, would need to be applied to the three-dimensional national soil matrix.

An integrated soil monitoring system could also include indicators of soil health and quality, of which carbon is only one indicator.  Monitoring soil C will not only be important from a reporting viewpoint, but could also feed into policy relating to land management.

Other systems were examined that considered the following:

  • the statistical uncertainty in estimates derived from a particular sample size

  • the sample size necessary to achieve a desired degree of accuracy

  • the cost of obtaining such a sample size

  • the limitations and assumptions inherent in the estimation, and

  • whether the result will answer the required question.

In particular, we considered:

  • National sampling of key cells or land-use categories. The cost of this sampling strategy at the 95% confidence level is estimated at ~$160,000 per year for field sampling alone.  It would be difficult under this system to attribute any change in soil C to a specific change in land-use activity and, according to the current IPCC guidelines, it would be difficult to determine whether such a change was anthropogenic

  • Periodic sampling of key cells or land-use categories. Such a system would be intensive and costly, and changes to soil C could take up to 40 years to reach steady state before the full contribution from land-use change would be available.

  • Updating national soil carbon estimates based on land-use change  The cost of such a system is mainly associated with updating the land-use database, an activity necessary for the forest and scrub monitoring system and therefore to some extent a sunk cost of the overall system.

4.2 Predicting national soil carbon from process-based “soil carbon” and associated models

Several models have been produced that can predict changes in the soil C pool.  They can provide a temporal dimension to changes in soil, and have generally been extensively validated using long-time-series data.  It is envisaged that they would complement and enhance the utility of the coefficients of change system under development.

4.3 Designing the monitoring system

The most robust strategy would be to have a combined system of direct sampling, modeling and using coefficients of change, by making the primary estimation of soil C change attributable to land-use change from historical and contemporary data, and to compare these estimates with those from paired sites, models, and multi-temporal sampling at benchmark sites.  The number of sites required will depend on the importance of the temporal component.  It is likely that at least a partial network of new sites will need to be established.

Based on the above, a number of studies were carried out to:

  • test the robustness of the three data layers

  • provide a temporal component to the coefficients of change (COC)

  • assess the contribution of benchmark sites to the monitoring system

  • test the capability of the system to explain the variance in soil C at the landscape scale and quantify the effects of land use and land-use change.

4.4 Testing the system

Three tests on the monitoring system were carried out during 1998/1999.

Benmore Range in Central Otago was the site used to compare soil C values predicted by the monitoring system with those determined by intensive sampling. The variable being considered was climate. Soil C was measured from 72 locations in the Range with various combinations of aspect, elevation, and precipitation. The total soil C derived by sampling and testing was 1.2 Mt C. The proposed monitoring system estimated a level of 1.6 Mt C, reasonably close to the value based on direct measurement considering the large difference in scales between these two approaches. The higher value determined by the monitoring system was considered to be due possibly to the lack of information on slope and aspect in the model.

Our soil C estimates were compared with those based on the FAO soil attribute and extent information. This uses FAO-UNESCO soil maps of the world produced at a scale of 1:5 000 000 for which New Zealand has been assigned 17 soil types.  The resulting analysis estimated 1207 Mt C in the 0–0.1 m layer and 1489 Mt C in the 0.1–0.3 m layer. This compares with the monitoring system estimates of 1208 Mt C and 1532 Mt C respectively.

The last test compares soil estimates from the monitoring system with those based on default IPCC soil C values, different soil/climate categories, and some land-use coefficients. Comparison of the default values with New Zealand mean soil C values for the same soil/climate categories show that New Zealand soils are generally higher in soil C stocks then global estimates (Tate et al. 1997b).

 

[ |