1. In 2011, the development of a set of “Sentinel Landscapes” was presented as “one of most innovative approaches proposed for CRP6” (2009 Stripe Review commissioned by the CGIAR Science Council).
As coordinator of the Borneo Sumatra Sentinel Landscape at the time of implementation, I would like to share with you not only the data or the way we obtained it, but more the process, the actual implementation and the challenges to built and manage such long-term research network
Sentinel Landscapes : the challenge of building long-term research networks
CIFOR Annual Meeting, October 2017

2. Several Regional Groups Robust governance structures
44 member networks
Several Regional Groups
Robust governance structures
Multiple partnerships at all level of organization (sites, countries, regions, global)
Thousands of papers published
Includes social sciences since 2005 (LTSER).
Slide 2 Background / LTSER networks   
This is not new…..Such initiatives started at the beginning of the 80s in US (LTER) boosted to international coverage in the 90s (ILTER), nowadays a 44 member networks with several Regional Groups and robust governance structures, multiple partnerships at all level of organization (sites, countries, regions, global), scientific knowledge reflected by ten thousands of papers published on the basis of data and findings generated at ILTER facilities and by ILTER teams. This includes social sciences since 2005, often referred now as LTSER.
Why this FTA Sentinel Landscape Network initiative then? Actually still a bit poor representation of developing countries in ILTER especially for Africa! In Africa only South Africa, Namibia and Malawi are members, in Latin America, Brasil, Mexico and Chile, in Asia Pacific, Malaysia, Philippines, Thailand, China, Japan; Indonesia, Laos, Cambodia, Vietnam are not members.
LTSER are known to be most useful in the monitoring of socio-ecological transitions.
Long-Term Socio-Ecological Research networks are known to be most useful in the monitoring of socio-ecological transitions.

3. Sentinel Landscape uses the concept of “Tree cover transitions” as unifying concept for livelihoods, landscape and governance
SL: A site or a network of sites, geographically or issue bounded, in which a broad range of biophysical, social, economic and political data are monitored, collected with consistent methods and interpreted over the long term.
A sentinel landscape is essentially a site or a network of sites, geographically or issue bounded in which a broad range of biophysical, social, economic and political data are monitored, collected with consistent methods and interpreted over the long term.
Site selection existing sites (current long term landscape scale sites or networks where CRP6 centers are already working);
The sentinel landscape network (SLN) is an initiative to set up long term socio-ecological research sites and to collect an integrated dataset (livelihood, institutions and environmental data) that support the strategic research framework of the CGIAR.
The ‘sentinel landscapes’ involves 200 research sites spread across 8 landscapes in 15 countries on 3 continents.
The 8 sentinel landscapes are Nicaragua and Honduras, Western Ghats (India), the Mekong (China, Laos), West Africa (Ghana-Burkina Faso), Western Amazon (Brazil, Peru, Bolivia), Borneo-Sumatra (Indonesia), CAFHUT (Cameroon) and Nile-Congo (Kenya, Rwanda, Democratic Republic of Congo (DRC)).

4. Key SL Research Questions (2012)
Is there a relationship between the variation in Tree cover/Tree quality and the variation of any of the four system level outcomes
reduction
in poverty
increased global food security
improvement of nutrition.
better management of
natural resources.
What explains spatial and temporal variation of tree cover?

5. Implementation and achievements
Year 1 and 2 ( )
Detailed analysis of existing networks and opportunities for collaboration ✖
Workshop to select sites (Nairobi, 2012) ✔
Creation of a working group on methods ✔
Develop partnerships with relevant partners ✔
Workshop method & data collection procedures (Ouagadougou, January 2013) ✔
Start carry out measurement campaign ✔

6. Implementation and achievements
Year 3 and 4 ( )
Method and Analysis workshop (Costa Rica, March 2014) ✔
Progress Meeting (Rome, Oct 2014) ✔
Workshop on institutional mapping (Montpellier, Dec 2014) ✖
Data collection implemented at each sentinel landscapes sites ( ) ✔
Outreach, World Forestry Congress 2015 ✔

7. Sentinel sites: Batang Lupar, Mentebah, Sarolangun, Merangin
Borneo-Sumatra SL: 4 selected sentinel sites, each measuring 10 x10 km2 , representing a variation in tree cover along the transition curve.
West Kalimantan
For each of the 8 sentinel landscapes, data was collected from 4 selected sentinel sites, each measuring 10x10 km2 , and representing a variation in tree cover over a 10 year time period.
Sentinel sites: Batang Lupar, Mentebah, Sarolangun, Merangin

8. Village level baselines and Household surveys IFRI instruments
A total of 1040 households surveyed between August and December 2014 from 40 villages.
Household demographics,
migration,
education,
asset ownership,
income sources,
household food security,
progress out of poverty,
crop production and sales, livestock products,
social networks, and
natural resource use.
Borneo – Sumatra Village level Baselines and Household surveys
A total of 1040 households were surveyed between August and December 2014 from 40 villages.
The data consists of information on household demographics, migration, education, asset ownership, income sources, household food security, progress out of poverty, crop production and sales, livestock products, participation in credit markets, social networks, and natural resource use.
In compliance with the CGIAR protocol on collecting sex-disaggregated data, approximately 50% of the respondents interviewed were women.

9. Land Degradation Surveillance Framework protocol
Land cover,
Tree and shrub densities and diversity,
Erosion prevalence,
Infiltration capacity, etc…
…. Analysis in Nairobi, infrared spectroscopy and wet chemistry.
LDSF protocol Indonesia
The LDSF (Land Degradation Surveillance Framework) carried out at the same four-100 km2
The LDSF is a stratified, randomized sampling design, developed to provide a framework for assessing processes of land degradation
Measured variables include: land cover, tree and shrub densities, tree diversity, erosion prevalence, infiltration capacity, along with an assessment of other soil characteristics
Soil samples were also collected (320 top (0-20 cm) and sub (20-50 cm) soil samples per site) and were first processed in Indonesia. Processed samples were shipped to Nairobi and subjected to infrared spectroscopy and wet chemistry analysis.

10. Re-measurement, changes since Year 2 and analyze the trends
Data repository ( )
Year 5 (2016)
Publication workshop (FLARE meeting Dec 2016) brainstorming on 8 to 9 papers using SL dataset.
Re-measurement, changes since Year 2 and analyze the trends

11. Weaknesses No analysis of existing ILTER networks
Partners not involved at the beginning
In most sites partners did not show much interest (or loose it quickly) = link with partners and donors not secured
Not same level of assistance for local teams between sites (some sites receive trainings, some not)
Data and feedback not given to the potential users at the end (even to us, cf. infrared spectroscopy)
Poor cooperation between FTA CG Centers, what’s more between CRPs!
Sometimes surprisingly poor level of science (remote sensing and tree diversity)
“Participatory” workshops
No real analysis of existing ILTSER
Partners not really involved at the beginning
In most sites partners did not show much interest (or loose it quickly) = link with partners and donors not secured
Not same level of assistance for local teams between sites (some sites receive trainings, some not)
Data and feedback not given to the potential users at the end
Poor cooperation between FTA CG Centers (competition) on this, cannot imagine between CRPs
Extremely basic and sometimes poor level of data (remote sensing and tree diversity weak)
“Participatory” workshops

12. More (much more) data to be included for socio ecological monitoring!

13. For consideration
Limitation in funding is the usual big constraint for LT research network, still…
Harmonization in methods and instrumentation needed
Already existing datasets, even not using the same methods should be considered (appropriate statistical tools that address differences in method)
Focus on the geographical gaps of ILTER
Set up regional group and partnership to join ILTER or ILTSER
Limitation in funding is the usual big constraint for LT research network, still…
Better harmonization in methods and instrumentation
More (much more) data to be included
Already existing datasets, even not using the same methods should be considered (appropriate statistical tools that address differences in method)
Focus on the geographical gaps of ILTER
Set up a structure to join ILTER or ILTSER

14. Thank you