1 of 10

Bridging micro- and macro-scale analysis

Justin Johnson and Uris Baldos

2 of 10

Agenda

  • Quick recap of key themes emerging from summer workshop
  • Specific challenges faced at the meso scale
  • A few first ideas on how to make progress
  • 30 minutes for group discussion on specific ways forward

3 of 10

Overview of meso-scale analysis

  • We have great experience on global/regional/national analysis
    • Trade policy, national decision making, price effects
  • We have good experience on landscape-scale (e.g. 10-300 meter grid-cell) conservation and planning
    • Which specific locations are most important to conserve for soil erosion, nutrient retention, etc.
  • Some recent work links these scales, but challenges abound
    • Several regional/local studies have done by researchers not yet used in “real-time” by policy makers (e.g. IEEM, GLOBIOM)
    • Global work is still in its infancy (e.g. GTAP InVEST) and could learn a lot from regional/local scale analysis especially in terms of crafting policies
      • Could be useful for identifying hotspots and identifying global boundary conditions for regional analysis

​

4 of 10

2nd GLASSNET Summer Workshop Series�Conducting Analysis Across Geographical Scales

  • Presentations and brainstorming on the newest Global-Local-Global applications that span multiple spatial scales.

​

  • Key conclusion: The meso-scale, i.e. the scale(s) between global/regional and land-scape level, are both extremely critical for our analysis but hard to model or understand

​

  • So far, key challenges identified include the following:
    • Greater resolution for socio-economic modeling of human behavior
    • Optimal precision for meso-scale analysis
    • Moving from single-activity to complex farming systems (e.g. mixed grazing and agroforestry)

5 of 10

Challenge #1 Greater resolution for socio-economic modeling of human behavior

  • We often focus on how policies affect production or consumption, but how do we move beyond these impacts?
    • Poverty, labor wages, payments to landowners, food security metrics such as caloric intake
  • Heterogenous decision making among agents is crucial, so far current economic models assume “representative” agent/sector/household at the national/regional level
    • Decision makers might not be market agents (politicians, other interests)
  • Household data may be useful to incorporate, but such surveys are based on samples that are nationally representative in the present, which may not hold in future decades

​

6 of 10

Example: Summarizing spatial heterogeneity via a calibrated function per region

6

Warm colors: least arable

​

Cool colors: most arable

7 of 10

Challenge #2 Optimal precision for meso-scale analysis

  • High-resolution grid-cell could provide different insights as compared to the case when using coarser grid-cells (e.g. InVEST carbon storage)
  • Using coarse data in fine scale analysis via downscaling can mean that the coarse data signal gets washed out as just some weighted average of biophysical inputs
  • But high-resolution outputs is not fully incorporated in economic analysis due to national-level modelling (i.e. aggregation of high-resolution results)
  • This ties back to greater resolution for economic modelling

8 of 10

Example: land-use change within GTAP-AEZ

9 of 10

Challenge #3 Moving from single-activity to complex farming systems

  • Most global economic models have very coarse resolution (i.e. national, regional) production systems

​

  • High-resolution agricultural models are available (e.g. GLOBIOM) but ignore interactions across agricultural activities in a grid-cell (e.g. mixed grazing and agroforestry)

​

  • Also, ignores landscape structure which is important for biodiversity and water quality issues

10 of 10

Possible ways to move forward in the near term

  • Challenge #1 Greater resolution for socio-economic modeling of human behavior
    • Explore specialized versions of economic models which generate additional distributional outcomes such as poverty and nutritional outcomes at the national-level
  • Challenge #2 Optimal precision for meso-scale analysis
    • Look for ways to increase relevant resolution in global economic models (greater AEZs) or use existing high-resolution agricultural models
  • Challenge #3 Moving from single-activity to complex farming systems
    • Add in to GTAP as alternate production activities?

​

  • In the long-run, high resolution economic data will be need to fully address these challenges