Reducing Climate Policy Risk: Improving Certainty and Accuracy in the U.S. Land Use, Land Use Change, and Forestry Greenhouse Gas Inventory

Climate change and climate policy are putting pressure on global landscapes like never before. To stay on track, we need 21st century systems to monitor land carbon sinks.

Blog by Seth Monteith, Tracy Johns, and Emily McGlynn

Report by Emily McGlynn, Kandice Harper, Serena Li, and Michael Berger

Climate change is putting pressure on global landscapes like never before. The U.N. Intergovernmental Panel on Climate Change (IPCC) indicates in a recent report that deforestation and intensive agriculture are responsible for nearly a quarter of human-caused global greenhouses gas emissions every year. This is accelerating climate-driven desertification, transforming large regions of Asia, Africa, and the Middle East. At the same time, policy makers are seeking to fight climate change by using land as a “carbon sink” to remove billions of tons of CO2 from the atmosphere in the coming decades.

For example, U.S. Senator Cory Booker recently filed a bill aiming to plant 16 billion trees on U.S. federal and private lands by 2050 and to restore 1.5 million acres of wetland. The hope is that these types of programs can bolster the U.S. carbon sink, which currently sequesters between 600-800 million metric tons (Mt) of CO2 every year into forests and soils, equivalent to offsetting the annual energy-related CO2 emissions of Texas, the largest emitter in the U.S. On a global scale, there have been calls to expand global forests by 900 million hectares, an area larger than China, while up to 700 million hectares could be required to supply bioenergy for carbon capture and storage.

Taken together, global landscapes will be pushed by climate impacts and pulled by climate policy in the coming decades on continental scales. Accurately tracking the implementation of these policies, including CO2 emitted and sequestered in landscapes, has become more important than ever to ensure the international community is meeting U.N. Paris Climate Agreement goals.

Yet, there is significant uncertainty in how much CO2 from the atmospheres is removed by landscapes. For example, the U.S. annual carbon sink estimates have an uncertainty range of roughly 580 Mt of CO2. This means that the amount of carbon sequestered in U.S. forests and soils, estimated last year at 714 Mt of CO2, might be as high as 1,078 or as low as 478, a range equivalent to the size of California’s energy-related emissions.

Figure 1 shows one way this uncertainty manifests – with continual updates to land carbon monitoring methods, each year the U.S. greenhouse gas inventory reports a completely different estimate of historical land carbon storage. Without significantly improving methods for measuring and monitoring land carbon, setting climate targets and tracking progress will become significantly more difficult.

Figure 1: The U.S. land use, land use change, and forestry carbon sink, as reported in the last ten years of U.S. greenhouse gas inventory reports.

Source: Ohrel, in press. Each line represents the historical data for a particular year’s inventory report. For example, the data contained in the 2019 inventory report is denoted in purple. The yellow indicates the range of data contained in reports from different years. Values for the years 2003 and 2004 are most varied with a difference of over 600 Mt CO2, with the 2009 inventory reporting a sink of nearly 1300 Mt CO2 while the 2019 inventory reports these years closer to 700 Mt CO2, for a difference of over 600 Mt CO2. Lines represent carbon stock change values only.

A new report has assessed where U.S. landscape carbon flux uncertainty comes from and how we might improve land carbon assessments going forward. The goal is to strengthen confidence in U.S. emissions reduction reporting while providing a blueprint for other countries. In Reducing Climate Policy Risk: Improving Certainty and Accuracy in the U.S. Land Use, Land Use Change, and Forestry Greenhouse Gas Inventory, experts identify, quantify, and recommend improvements for over 90 elements of uncertainty that together make up around 70% of the uncertainty in the economy-wide U.S. greenhouse gas inventory (Figure 2). They also provide initial estimates for omitted greenhouse gas fluxes, those categories left out of the inventory due to data limitations (Figure 3).

Figure 2: Sources of uncertainty in the land use, land use change, and forestry greenhouse gas inventory.

The number next to each element represents its contribution to the uncertainty (95% confidence interval) of its respective calculation in Mt of CO2 emitted. The project team analyzed over 90 uncertainty elements across input datasets, models, and calculation methods for greenhouse gas fluxes in forests, croplands, grassland, settlements, and wetlands, using statistical and survey methods. It is not valid to add together all the uncertainty attribution values to find total uncertainty of the land use, land use change, and forestry greenhouse gas inventory.

Figure 3: Omitted greenhouse gas fluxes in land use, land use change, and forestry greenhouse gas inventory.

In addition to estimating sources of uncertainty, the report also identifies omitted greenhouse gas categories. The U.S. greenhouse gas inventory already recognizes most of these gaps and has listed many of them as planned improvements to be incorporated into future inventories. The report takes this recognition one step further by using literature, activity data, and IPCC and U.S.-specific emission factors to provide rough estimates of omitted greenhouse gas fluxes. These estimates are meant to help set priorities for addressing omitted fluxes. The magnitude of each flux estimated here is likely to have low accuracy, due to lack of data. Note that “PRP” refers to paddock, rangeland, and pasture manure deposition, the only source of N2O emissions on federal grasslands currently included in the land use, land use change, and forestry greenhouse gas inventory.

The largest sources of uncertainty include:

  • A need for more forest biomass measurements: Estimating national forest carbon by scaling up sample plot measurements contributes over one-third of the uncertainty identified in the report.
  • Soil carbon and nitrogen modeling: DayCent, a complex biogeochemical model, contributes nearly another third of uncertainty, and requires more data and model structure development to estimate carbon storage, methane, and nitrous oxide emissions from cropland and grassland soils.
  • Lack of representative tree carbon data: Using values averaged over many tree species, ages, and climates to convert tree height and diameter into carbon represents over 10% of uncertainty.


The largest omitted greenhouse gas fluxes include grassland and wetland soil carbon in Alaska, which could emit 90 Mt of CO2 annually, reducing the national land carbon sink by 12%. Emissions from urban soils and federal lands were also identified as significant omissions.

Addressing major sources of uncertainty and omitted fluxes will require significant investment, research, and coordination across federal agencies and experts. The report outlines key recommendations, including better use of remote sensing data to measure forest biomass and more extensive collection of soil carbon measurements. In the longer term, the report recommends that the U.S. develop a nationally-consistent land inventory system that calculates carbon across forests, croplands, grasslands, urban areas, and wetlands using the same data sources and models – overcoming the challenges of the current system’s medley of methods across land types.

To learn more about land carbon uncertainty and recommendations to address it, read the full report – written by Emily McGlynn, Kandice Harper, Serena Li, and Michael Berger – here.