Biophysical Modelling (Levels 1 and 2) Thanks for inviting me and for coming today! […who am I? what is my role?] The project is a collaboration of the United National Statistics Division, United Nations Environment Programme, the Secretariat of the Convention on Biological Diversity It is funded by the Norwegian Ministry of Foreign Affairs. This module is on Biophysical Modelling…

Contents Review of Level 0 (5m) Level 1 (Compilation) Learning objectives Presentation (15m) Group exercise (15m) Level 2 (Data providers) Closing discussion (15m) SEEA-EEA: Biophysical modelling

Review of Level 0: Biophysical modelling SEEA-EEA: Biophysical modelling

Tools 3: Biophysical modelling Biophysical modelling is also an important tool…It is used in many of the accounts, as we have seen. SEEA-EEA: Biophysical modelling

Level 0: Tools 3: Biophysical modelling What? Four main approaches: Why? Estimate Ecosystem Services across spatial units and time Estimate Ecosystem Capacity from Ecosystem Condition Combine data from various sources and scales (e.g., point field data and satellite data) Estimate unknown data values GIS-based spatial modelling approaches have methods built-in Look-up tables Statistical approaches Geostatistical interpolation Process-based modelling (Conclusion) We will discuss each of the four main approaches in turn. SEEA-EEA: Biophysical modelling

Level 0: Tools 3: Biophysical modelling Approaches: Look-up tables Statistical approaches Geostatistical interpolation Process-based modeling Attribute values for an ecosystem service (or other measure) to every Spatial Unit in the same class (e.g., a land cover class). Example: Benefits Transfer one ha of forest = $5000  attribute to each ha of forest error rate: 60-70% In this example, we assume that the monetary value of a hectare of forest is $5,000. We attribute this value to each hectare of forest. Note that the approach has a high error rate since different locations will have different values. Lookup tables can be used for more than ecosystem values (per hectare). For example, you can attribute a carbon storage class to land cover classes. SEEA-EEA: Biophysical modelling

Level 0: Tools 3: Biophysical modelling Approaches: Look-up tables Statistical approaches Geostatistical interpolation Process-based modeling Estimate ecosystem services, asset or condition based on known explanatory variables such as soils, land cover, climate, distance from a road, etc., using a statistical relation. Example: Function Transfer Value = f(land cover, population, roads, climate) Error rate = 40-50% Statistical approaches refers to using a known statistical relationship to estimate unknown values. This usually requires developing a transfer function using regression. In this example, we estimate the “value” based on land cover, population, roads and climate. This reduces the error rate somewhat, since we’re using more information than simply the land cover type. SEEA-EEA: Biophysical modelling

Level 0: Tools 3: Biophysical modelling Approaches: Look-up tables Statistical approaches Geostatistical interpolation Process-based modeling Use algorithms to predict the measure of unknown locations on the basis of measures of nearby known measures: Example: Kriging Error rate = ? Known Unknown In this case, we can estimate the value for the red dot, based on the values of nearby blue dots. The error rate depends on the context and the variability of the measure we are estimating. [Example actually shows a simpler approach: Inverse Distance Weighting] SEEA-EEA: Biophysical modelling

Level 0: Tools 3: Biophysical modelling Approaches: Look-up tables Statistical approaches Geostatistical interpolation Process-based modeling Predict ecosystem services based on a set of future condition or management scenarios: Example: Scenario for future services based on expected changes in land cover, demand and management Error rate = 100% Many models use a combination of approaches to estimate future conditions. For example, if land use and population are expected to change over 10 years, what are the flows of services? The error rate is high, since it is assuming future conditions, which are difficult to predict. SEEA-EEA: Biophysical modelling

Level 1: Biophysical modelling SEEA-EEA: Biophysical modelling

Level 1: Learning Objectives Understand the role of spatial and of temporal modelling Required prior knowledge: the concept of ecosystem services SEEA-EEA: Biophysical modelling

Level 1: Tools 3: Biophysical modelling Biophysical modelling for the purpose of ecosystem accounting is required to: Provide estimates when no data are available Provide estimates of future Conditions, Capacity and Services Supply Provide links between Extent, Conditions, Capacity and Services Supply It does so by: Capturing the spatial variability of ecosystems Capturing the temporal variability of ecosystems Modelling: make it as simple as possible, but not simpler (after Einstein) SEEA-EEA: Biophysical modelling

Spatial variability in ecosystems Ecosystems vary across space because of differences in altitude, soils, vegetation, climate, etc. and because of spatial differences in the management of ecosystems (e.g. as a function of distance to road or settlement) The Wetland ecosystem ‘De Wieden’, the Netherlands consists of different zones with different vegetation, humidity, soils, etc. (Source: Hein et al., 2006) SEEA-EEA: Biophysical modelling

Modelling spatial variability Modelling spatial variability is required because ecosystem accounting aims to record flows of ecosystems and stocks of ecosystem capital; and both flows and stocks can be spatially very heterogeneous. Flows and stock modelling usually requires the combination of spatial datasets (maps) and point-data (e.g. from surveys). To combine these datasets, and to understand ecosystem flows and assets in areas with no data, a spatial model can be used. SEEA-EEA: Biophysical modelling

Modelling temporal variability Temporal models are required to analyse ecosystem assets The Net Present Value of an ecosystem asset is the sum of the net, current and discounted future values generated by the ecosystem These values depend upon the Capacity of the ecosystem to regenerate, for instance the regrowth of the forest after harvesting. Temporal models reveal the regeneration/regrowth of the ecosystem The increase in a biological resource often follows an S-curve, where growth levels approach zero of at the carrying capacity. SEEA-EEA: Biophysical modelling

Level 1: Tools 3: Biophysical Modelling Compilation group exercise (15m) Situation Have 5 years of historical data on timber volume, harvest, fire and storm damage Need to model regrowth, opening and closing stock Need to estimate “sustainable” harvest Objective (Groups of 3-5 persons) Calculate regrowth, closing stock, opening stock for each year Calculate “sustainable” harvest for Year 6 and onwards Please see the exercise sheet: SEEA-EEA: Biophysical modelling

Level 1: Tools 3: Biophysical Modelling Step 1: Calculate regrowth, opening stock and closing stock for years 2 to 5 Opening Stock for Year 2 is Closing Stock for Year 1 Regrowth is 5% of current year Opening Stock SEEA-EEA: Biophysical modelling

Level 1: Tools 3: Biophysical Modelling Step 2: Calculate Average Harvest (Year 1 to Year 5) (B) Average Fire and Storm Damage (Year 1 to Year 5) (A) Expected “Year 6” Additions (5% of Year 5 Closing Stock) Recommended “sustainable” harvest ((A) – (B)) SEEA-EEA: Biophysical modelling

Level 1: Tools 3: Biophysical Modelling Is everyone clear on the objectives? 15 minutes group work Please ask questions Results: Report Recommended “sustainable” harvest What could you add to the account to improve the estimate? SEEA-EEA: Biophysical modelling

Level 1: Tools 3: Biophysical Modelling The Answers A harvest rate of 28.0 m3/year would maintain the closing stock (912.9 m3) Could include: replanting, different species, age/size growth rates… SEEA-EEA: Biophysical modelling

Level 2: Biophysical modelling SEEA-EEA: Biophysical modelling

Level 2: Tools 3: Biophysical modelling Learning objective: To understand which types of modelling approaches can be used for the spatial and temporal modelling of ecosystem services in an accounting context SEEA-EEA: Biophysical modelling

Spatial modelling To estimate ecosystem services flows and ecosystem assets across spatial units in the landscape In spatial models: ecosystem condition indicators, ecosystem services flow and ecosystem assets are defined and analysed for each spatial unit This requires Geographical Information Systems (GIS), where the spatial unit is usually a pixel (i.e. a grid cell in a grid, with specific x and y coordinates) SEEA-EEA: Biophysical modelling

Four types of spatial models Look-up tables: specific values for an ecosystem service or other variable are attributed to every pixel in a certain class, usually a land cover class.   Statistical approaches: relate ecosystem services flow, asset or condition to explanatory variables such as soils, land cover, climate, distance from a road, etc., using a statistical relation derived from survey data.   Geostatistical interpolation: techniques such as kriging rely on statistical algorithms to predict the value of un-sampled pixels on the basis of nearby pixels in combination with other characteristics of the pixel. Process based modeling: involves predicting ecosystem services flows based on a set of environmental properties, management variables and/or other spatial data sources, based on modelling of the ecological and/or ecosystem management processes involved. (We will have seen this in Level 0) SEEA-EEA: Biophysical modelling

Ecosystem services Central Kalimantan Model used Look Up Tables (every land cover class is attributed a specific carbon storage value) Kriging (values are interpolated from samples) Carbon storage Source: Sumarga and Hein, 2014 Timber production In these examples, estimates of Ecosystem Services were made using different Biophysical Modelling approaches, using different Condition measures. SEEA-EEA: Biophysical modelling

Ecosystem services Central Kalimantan Orangutan habitat Model used: Statistical model (Maxent) (habitat suitability predicted on the basis of forest cover, distance from road, etc.) Process-based Model (primary ecosystem production minus soil respiration) Carbon sequestration Source: Sumarga and Hein, 2014 SEEA-EEA: Biophysical modelling

Modelling ecosystem services Can be done in stand-alone GIS packages such as ArcGIS or Q-GIS (freeware) or Modelling packages such as ARIES or InVEST Advantages of ARIES and InVEST is that some modules are pre-defined; Advantage of standard packages is that they are fully flexible and not dependent upon (knowledge of) a specific modelling program SEEA-EEA: Biophysical modelling

Question 1 Which four types of spatial models can be distinguished? What is the differences between them? Explain in your own words. SEEA-EEA: Biophysical modelling

Answers 1: Four types of spatial models: Look-up tables: specific values are attributed to every pixel in a certain class, usually a land cover class.   Statistical approaches: ecosystem services flow, asset or condition is related to explanatory variables such as soils, land cover, climate, distance form a road, etc., using a statistical relation derived from survey data.   Geostatistical interpolation: techniques such as kriging rely on statistical algorithms to predict the value of un-sampled pixels on the basis of nearby pixels in combination with other characteristics of the pixel.   Process based modeling: involves predicting ecosystem services flows based on modelling of the ecological and/or ecosystem management processes involved. SEEA-EEA: Biophysical modelling

Question 2 Select one or two ecosystem services relevant for your case study area of interest Which different sets of data are required to model and map this service? How can these datasets be combined? Possible answers: Services  ecosystem types  National data: Food security, Crops  agricultural land  agricultural surveys, soil condition, climate, erosion, practices (fertilizer, pesticide) Timber  forested areas  forest inventories, species, regeneration rates, harvest rates, storm and fire rates Biomass for energy (forested areas, agricultural land)  see above Carbon sequestration (forested areas, mangroves, grasslands)  land cover types, vegetation types, Flood control (wetlands, mountain forests, upper catchment areas)  vegetation types and density, slope, soil type Water purification (wetlands, upper catchment areas)  vegetation types, flow rates, wetland area Recreation (forests, wetlands, beaches)  species, diversity, condition SEEA-EEA: Biophysical modelling

Level 2: Tools 3: Biophysical Modelling Discussion and questions Take home points Biophysical modelling can help fill data gaps Biophysical modelling can help estimate future conditions, services and capacity Many biophysical models are spatial and combine data from many sources Geographic Information Systems (GIS) and pre-defined modelling packages have methods and formulas included SEEA-EEA: Biophysical modelling

Level 2: Biophysical Modelling References Hein, L., Van Koppen, K., De Groot, R. S., & Van Ierland, E. C. (2006). Spatial scales, stakeholders and the valuation of ecosystem services. Ecological Economics, 57(2), 209-228. Sumarga, E. And Hein, L., 2014. Mapping Ecosystem Services for Land Use Planning, the Case of Central Kalimantan. Environmental management, pp. 1-14. Further Information SEEA Experimental Ecosystem Accounting (2012) SEEA-EEA Technical Guidance (forthcoming) Detailed supporting document on “Biophysical Modelling and Analysis of Ecosystem Services in an Ecosystem Accounting Context” by Lars Hein SEEA-EEA: Biophysical modelling

Acknowledgements Materials prepared by: Adapted from: Michael Bordt Regional Adviser on Environment Statistics ESCAP Statistics Division bordt@un.org Adapted from: Lars Hein and Michael Bordt Advancing Natural Capital Accounting, a collaboration between The United Nations Statistics Division (UNSD), United Nations Environment Programme (UNEP) and the Secretariat of the Convention on Biological Diversity (CBD) and is supported by the Government of Norway. https://unstats.un.org/unsd/envaccounting/eea_project/default.asp Contact: seea@un.org SEEA-EEA: Biophysical modelling 33