Harmonised observations of climate forcing across Africa: an assessment of existing approaches and their applicability

An inventory of methodological protocols was compiled between November 2017 and October 2018 to obtain a comprehensive overview of the methodologies and techniques that have been developed and used over the last two decades by the international and European research community to measure and estimate environmental variables across the terrestrial, atmospheric and oceanic domains. The point of departure for the selection of relevant protocols was a set of 58 environmental variables (see figure 1) which had previously been identified in line with the SEACRIFOG initiative to be essential for the quantification of the main climate forcing components in Africa (Beck et al2018, López-Ballesteros et al2018a). To be considered, each protocol had to be directly or indirectly related to the measurement of at least one of the essential variables. Regarding the date of publication, we only considered protocols published after 2000 to assure that these are reasonably up to date, hence collated protocols were published between 2004 and 2018, with the majority (∼80%) published within the last four years (2014–2018). Protocols were identified using a keyword search on the web-portals of international and regional research agencies, networks and projects related to environmental monitoring as well as through expert consultation within the SEACRIFOG consortium. Only open-access protocols/materials were considered. Several metadata attributes were collected for each protocol in order to characterise its content in a systematic way (see metadata glossary in table 1). The entire protocol inventory metadata is publicly available via the web-based 'SEACRIFOG Collaborative Inventory Tool' at https://seacrifog-tool.sasscal.org. In addition to information on the protocols, the tool contains information about SEACRIFOG essential variables as well as corresponding observation infrastructures and data products relevant to the African continent.

Zoom In Zoom Out Reset image size

A preliminary assessment was performed in order to evaluate the feasibility of the reviewed protocols based on the following criteria:

• equipment costs (C_eq; i.e. instrumentation and other experimentation material costs)²
• operational costs (C_op; i.e. calibration and maintenance of equipment and experiments, annually)³
• installation effort (E_i; i.e. number of person*hours to setup the station or experiment)
• maintenance effort (E_m; i.e. number of person*hours related to sampling frequency, in case of manual measurements, or routine repairs/calibrations in case of automated measurements)
• knowledge needed (K; none, low, medium or high)
• measurement mode (manual or automatic)

For this assessment, we considered only those protocols from the inventory that deal with ground-based and sea-borne observations (i.e. in situ), which equates to 82 out of the 140 protocols within the inventory. Although satellite data are complementary to in situ observations and essential for model building, spatial and temporal upscaling and even site selection, protocols related to remote sensing were excluded from the analysis since associated direct costs can be hardly derived from their respective documents. Likewise, data processing and/or activity-based emission reporting protocols were not considered in the analysis. Likewise, strategic/implementation plans from international RIs (e.g. WMO, GAW, WIGOS) and protocols reviewing many methodologies simultaneously or specifying data requirements (e.g. WMO ECVs requirements) were also excluded from the analysis because their implementation costs cannot be directly derived from the respective documents.

These criteria have been previously used by Halbritter et al (2020) and collectively agreed by the researchers consortium involved in the ES1308 ClimMani COST Action⁴ (Climate Change Manipulation Experiments in Terrestrial Ecosystems—Networking and Outreach). Generally, the assessment of the knowledge, financial and human resources required for the implementation of a protocol were derived from the information derived from the protocols reviewed, the research experience of the co-authors and from expert consultations within the SEACRIFOG consortium. The categories used to classify the protocols are relatively coarse because the related information is not available for all protocols and the resource requirements are to a certain extent context-dependent as stated by Halbritter et al (2020). The feasibility of protocol implementation was rated by using the categories presented in table 2 below. In those cases where there are several station classes (i.e. tier 1 and 2 stations), the feasibility assessment was performed on the station type with the minimum cost.

Additionally, we derived five indices in order to be able to judge each protocol's feasibility. Certainly, the construction of such indices is a largely arbitrary exercise, which nevertheless adds value in interpreting and ranking the protocols for a given context. For the purpose of this assessment, we build basic indices based on simple addition of the respective underlying indicators (table 2).

To calculate these indices, ordinal rating values ('None', 'Low', 'Medium', 'High') and cost ranges ('None', '€', '€€', '€€€') were firstly translated into numerical values (0, 1, 2, 3, respectively). The indicator 'Knowledge needed' received a double weighting as compared to others, since we consider knowledge an asset that requires both capital and human effort over extended periods. Afterwards, indices were calculated as follows:

$$\begin{equation*}Initial\ Investment\ Index{{ }}\left( {II} \right){{ }} = {{ }}{C_{eq}} + {{ }}{E_i}\end{equation*}$$

$$\begin{equation*}Operational\ Investment\ Index{{ }}\left( {OI} \right){{ }} = {{ }}{C_{op}} + {{ }}{E_m}\end{equation*}$$

$$\begin{equation*}Financial\ Costs\ Index{{ }}\left( {FC} \right){{ }} = {{ }}{C_{eq}} + {{ }}{C_{op}}\end{equation*}$$

$$\begin{equation*}Human\ Capacity\ Index{{ }}\left( {HC} \right){{ }} = {{ }}{E_i} + {{ }}{E_m} + {{ }}2 \times K\end{equation*}$$

$$\begin{equation*}Feasibility\ Index{{ }}\left( {FI} \right){{ }} {=} {{ }}Max\ Value {-} \left( {II{{ }} {+} {{ }}OI{{ }} {+} {{ }}2 \times K} \right)\end{equation*}$$

The values for all indices were translated to a 0–10 scale. For II, OI, FC and HC, low values correspond to low financial and human resource needs for the implementation of a given protocol. When considering the feasibility index FI, low FI values correspond to low feasibility (i.e. high financial and human resources needs and accordingly high values for the other indices).

The protocol inventory is publicly accessible at https://seacrifog-tool.sasscal.org and currently contains metadata for a total of 140 methodological documents in different formats such as guides, protocols, technical reports, databases, data sheets, software, code, and models. A synthesised inventory version is provided in table S1 (stacks.iop.org/ERL/15/075003/mmedia) of supplementary material to this publication.

Figure 1 illustrates the number of protocols in the inventory associated with each variable considered essential for the quantification of climate forcing across the African continent. Since a given protocol can be linked to more than one variable, the number of protocol-variable links (615 in total) indicated in figure 1 exceeds the number of protocols in the inventory (140). As more than half of the essential variables are related to the terrestrial domain, protocols linked to terrestrial variables dominate the inventory (73%) compared to the atmospheric (18%) and oceanic domains (9%).

For the purpose of this study, we considered the main application of a protocol, distinguishing between observation (ground-based, sea- and space-borne), data processing and activity-based emission reporting. In total, 77% of protocols in the inventory have observation (ground-based, sea-borne, space-borne or combinations thereof) as the main application, whereas 14% are focused on data processing and 9% refer to emission reporting approaches. As illustrated in figure 1, there are numerous protocols guiding observation as well as data processing for variables in the terrestrial domain. For variables in the atmospheric and oceanic domains, hardly any data processing protocols could be identified. Finally, protocols guiding emission reporting were identified for the main GHGs as well as for atmospheric concentrations of important precursors and other climate-relevant trace gases.

In terms of geographic scope, a third (47) of the protocols are considered applicable anywhere around the globe. These were developed by international institutions (e.g. CGIAR, FAO, GEOBON, ICRAF, IPCC, UNFCCC or WMO), global research networks focused on environmental monitoring (e.g. FLUXNET), global manipulation experiments (e.g. Drought-net, ClimMani COST Action) and other international projects or initiatives that have been completed (e.g. IOCCP or RECCAP). Another 19 (14%) protocols are also applicable globally, but refer to specific biomes such as tropical forests, mangroves, seagrass meadows, urban areas, etc (see table S1 in supplementary material for detailed information). The remainder of the protocols in the inventory (74 protocols; 53%) were developed with applicability for a certain world region only (e.g. ICOS protocols for the European continent). Only 11% (15) of the identified protocols were developed specifically to be applied on the African continent (e.g. the protocols developed by VitalSigns). However, note that protocols developed for other world regions can be adopted in the African context in line with a harmonisation approach (except those which are designed for a region-specific land cover or land use).

Additionally, 94% of the organisations that developed the protocols are long-term initiatives (i.e. sustainably funded institutions such as WMO, ICOS, etc), while the rest of the protocols were developed by projects (i.e. short-term initiatives). The latter generally coincide with completed project outputs that represent a collective effort envisaged to advance in the methodological harmonisation within a specific research field (e.g. the International Ocean Carbon Coordination Project—IOCCP).

An overview of the links between the publishing institutions and both the earth system domains as well as the main application of the corresponding protocols is provided in figure 2. The purpose is to guide researchers interested in a specific domain and/or type of activity to the corresponding institutions and their protocols. For example, researchers interested in sea-borne observations are referred to a small number of specialized (and mainly project-based) initiatives (e.g. FixO3), whereas practitioners interested in emission reporting are rather referred to international institutions such as IPCC, UNFCCC or FAO.

Zoom In Zoom Out Reset image size

In our inventory, protocols primarily related to the atmospheric domain were developed by 12 different initiatives including international monitoring networks associated with international environmental agencies such as WMO and other initiatives at regional level (e.g. ICOS). In the oceanic domain, the role of individual projects to support the harmonisation of international oceanic observations through the development of guidelines on methodological and data-related best practices is substantial. Note that an extensive compendium of methodological material for oceanic observations is provided by the open-access catalogue of ocean best practices⁵. This resource, however, is not included in the present inventory as it covers disciplines and research objectives that go beyond the focus of the present study on the estimation of climate forcing. Protocols primarily related to the terrestrial domain, which for the purpose of this assessment includes freshwater systems, salt marshes and mangroves, were developed by a large number of initiatives, the ones with the most protocols being ICOS, ICP Forests and RAINFOR/AfriTRON. An important aspect of the terrestrial domain is the presence of both natural and managed systems. Protocols for the measurement of variables related to agricultural and livestock systems are those developed by FAO, CGIAR-CCAFS, GHG Protocol, Vitalsigns and INRA-RMT-ADEME. The latter is a complete/comprehensive guide that include several methods used to measure GHG, ammonia and nitrous oxides emissions from livestock farming.

A total number of 82 protocols dealing with ground-based and sea-borne observations were selected from the protocols inventory to assess the feasibility of their implementation. The results of this analysis are shown in table 3, where Initial Investment Index (II), Operational Investment Index (OI), Financial Costs Index (FC), Human Capacity Index (HC) and the overall Feasibility Index (FI) were calculated for each protocol by using approximate estimates of its equipment and operational costs, installation and maintenance efforts, and the knowledge needed for its implementation (see table S2 for detailed information).

Protocols that were found to be relatively easily feasible (FI > 6) include instructions for station/experiments set up, characterisation of site vegetation (developed by ICOS-ETC) and visual assessment of tropical vegetation status (developed by RAINFOR), because both financial costs and human capacity needed for their implementation are relatively low. Additionally, some other protocols focused on yield estimations, soil sampling and land cover/use validation in agro-systems (developed by VitalSigns) are considered highly feasible because they do not require either sophisticated instrumentation or highly trained personnel. Likewise, protocols describing precipitation and albedo measurements and approaches to estimate presence of soil microbes and animals, wet nitrogen deposition and atmospheric particles were also identified to be highly feasible due to their affordable equipment and easy implementation. For almost half of the assessed protocols the feasibility index is within an intermediate range (4 < FI < 6), which means that feasibility of implementation of a given protocol depends strongly on the specific context and purpose of the observation of interest (often high cost vs. low maintenance or vice versa). Many of the corresponding protocols are focused on meteorological (e.g. temperature, radiation, wind), vegetation (e.g. aboveground and litter biomass, wood density, leaves, needles, roots) and soil (e.g. carbon, nitrogen, nutrients, soil water content) measurements, while fewer deal with the measurement of greenhouse gas concentrations and fluxes in the atmosphere and in terrestrial and oceanic domains. Protocols with a low feasibility index (FI < 4) correspond to those for which both cost and human capacity needed for implementation are high. These include most of the guidelines developed to perform measurements in the oceanic systems (e.g. FixO3 guidelines) together with protocols describing complete meteorological stations (e.g. ICP Forests protocol), and approaches to directly measure biosphere-atmosphere GHG exchanges in natural and agricultural systems (e.g. ICOS-ETC and Global Research Alliance protocols).