In the third session, steering committee member Henry Harpending (University of Utah) chaired a session that included presentations by Stephen Polasky (University of Minnesota), Siwa Msangi (International Food Policy Research Institute), and James (Jae) Edmonds (Pacific Northwest National Laboratory), followed by discussion with participants.
Stephen Polasky, Fesler-Lampert Professor of Ecological/Environmental Economics and Regent’s Professor, Department of Applied Economics, University of Minnesota
Stephen Polasky noted that the previous sessions focused on population, whereas his session focused on the links between population, human actions, and ecosystems (including variables such as land use, water use, and land management), as well as how those variables feed into human well-being. He used the term “ecosystem services” to describe the benefits the ecosystem provides to people. Polasky then identified two challenges to society: (1) economic development to alleviate poverty and malnourishment, and (2) conservation of natural capital to maintain biodiversity and ensure that humanity fits into the Earth’s system.
Polasky stated that to support 10 billion people in 2050, the world gross domestic product (GDP) would need to increase by a factor of eight.
(This number also assumes an overall increase in affluence in that time period.) In comparison, the world economy grew by a factor of 40 from 1900 to 2000. However, Polasky reasoned that, in today’s world, attaining an eight-fold increase would be difficult and questioned if, while building up economic capital in the form of manufactured and human capital, humans are simultaneously eroding natural capital by not sufficiently considering environmental issues. He defined “natural capital” as the natural assets that provide ecosystem services, possibly multiple ecosystem services, such as a forest (natural capital) that provides ecosystem services (timber production, carbon storage, water regulation, habitat, and tourism). According to the Millennium Ecosystem Assessment (2005), which considered the status of and trends in biodiversity, ecosystems, and biodiversity are both essential for human well-being, and the ecosystem-services concept provides a central theme and organizing principle. The assessment looked at each ecosystem service and evaluated whether it was increasing or decreasing over time. In general, services related to food (such as livestock and food production) are increasing, but all other ecosystem services show a decreasing or level trend. Thus, according to the assessment, while living standards are increasing around the world, natural capital is on the decline (Millennium Ecosystem Assessment, 2005).
Polasky then discussed biodiversity in more detail. The Millennium Ecosystem Assessment documented that the natural historic rate of species extinction is between 0.1 and 1.0 per thousand species per millennium; recent extinction rates, however, are two to three orders of magnitude higher, and future rates are projected to be higher still. These numbers are based on models, not observations, resulting in a large error range (Millennium Ecosystem Assessment, 2005). He proposed that biodiversity and ecosystems have a symbiotic but strained relationship. Biodiversity, he explained, is an attribute, and while it contributes to ecosystem services, it is not an ecosystem service in and of itself. During the Millennium Ecosystem Assessment development, there was some tension in the community about the increased use of ecosystem services over biodiversity, as maintenance of biodiversity can be seen as a goal of its own. Most drivers of change in both ecosystem services and biodiversity are based in human activity, such as changes in land use, climate, and the nutrient cycle; pollution; and the movement of biota (i.e., invasive species).
Analyzing one of those human activities—land use—in more detail, Polasky divided land use history into five stages: pre-settlement, frontier, subsistence, intensifying, and intensive. The intensifying period was marked by huge changes, as the world’s ecosystem use transitioned to intensive agriculture and urban areas. He pointed out that people often use the word “loss” when describing this land use transition (harking back to Turner’s description of the Cassandra viewpoint), but, more neu-
trally, humans should be thought of as change agents, converting habitat from one form to another. He said a debate has emerged about peak land use—will expanding the land use base result in more conversion of habitat, or has peak conversion already been reached? He showed that crop land use in the United States declined from 1950 to 2000, as increases in yield offset increases in demand. Technology has been the dominant force in land use transition in the United States recently, but Polasky cautioned the trend was not likely to continue domestically.
Polasky explained that ecosystem services are what economists might call “public goods”; in other words, it is difficult for private companies or landowners to receive payment or reward for the natural benefits of their land. As a result, he said, landowners have little incentive to maintain the natural capital of their land. For example, adding nitrogen and phosphorus is of personal benefit to a farmer, but not to the overall natural resources of the land and the surrounding and downstream areas. Maintaining ecosystems services, Polasky said, has received significant interest recently among scientists and governments (for example, in the United States and China), and it will require a multidisciplinary effort to do so. He identified three tasks to mainstream the maintenance of ecosystem services: (1) understand the provision from the ecosystem, (2) understand the value (such as creating benefits and affecting human well-being), and (3) create incentives via policy.
Polasky then provided an extensive example of land use in a study of the Willamette Basin in Oregon, which looked at the impact of alternate land use patterns on biodiversity and ecosystem services (Polasky et al., 2008). The study combined a biological model (focusing on terrestrial species) with an economic model (focusing on the value of agriculture, timber, and housing development). The study considered nine different land use alternatives to maximize species as a function of land for a given economic score—in other words, the study attempted to assess the best one could do for biodiversity from the landscape. The study found a steep curve: Species conservation can be increased at little cost initially, but it becomes increasingly expensive later in the development phase. The study showed how well it is possible to perform if multiple benefits from the landscape are taken into account.
A second study of the same area (Nelson et al., 2009) considered multiple ecosystem services and trade-offs among them, including services such as water quality, flood control, soil conservation, carbon sequestration, species conservation, and market commodities. Rather than solve a multidimensional optimization problem, the study worked with groups in the Willamette Basin to predict likely changes through time to develop three possible future scenarios: plan trend (or “business as usual”), development, and conservation. Polasky posited that the relative ranking of
the scenarios would depend on the set of ecosystem services most valued; however, the conservation scenario showed the most improvement across the scenarios for all ecosystem services outputs, with the exception of market value. He suggested that if the value of certain ecosystem services were included—for instance, if landowners were compensated for carbon sequestration—there would be no significant trade-offs across the different scenarios, and the conservation scenario would become more attractive. The failure to incorporate the value of ecosystem services and biodiversity in planning, however, results in poor outcomes. Polasky postulated that the real challenge is one of integrated thinking—simultaneously integrating development and conservation to both spur economic development and enhance natural capital.
Polasky referred to his earlier statement that GDP needs to experience an eight-fold increase to support 10 billion people by 2050, noting that simply scaling up “business as usual” cannot produce enough for a population of 10 billion. He called for improved technology (better output per unit input), as well as improved institutions and incentives to increase output of goods and services without a negative impact on natural capital. He noted economists frequently write about the importance of incentives, paraphrasing a 2011 Science article (Kinzig et al., 2011) that “you get what you pay for and you don’t get what you don’t pay for.” Because ecosystem services are currently supplied for free and without reward, they may soon not be supplied at all. Returning to the list in the Millennium Ecosystem Assessment (2005) that shows the change in various ecosystem services through time, Polasky noted a clear and systematic distinction between those services that increased and those that declined: the increased services (all in food production) are compensated for that production, while those who control the declining services do not receive a tangible benefit for their production. He stated that although a full understanding of the system dynamics at play is not in hand, enough is known to improve on current performance and characterized the integration of decisions about natural capital into societal decision making as a good beginning.
Siwa Msangi, Senior Research Fellow,
International Food Policy Research Institute
Siwa Msangi began his presentation by stating that there is a pressing need for additional provisioning of food to preserve nutrition and well-being as the Earth’s population expands. He said that the challenge
presented by climate change—specifically, the impact of climate change on water availability—affects the ability to meet current and future food production needs. Water scarcity, one of the greatest challenges associated with climate change, is already a widespread problem, encompassing both quantity (how much water is available) and timing (when the water is available). With increasing global temperatures, crop evapotranspiration requirements are likely to increase. He postulated that it is important to understand the role of investments in maintenance and management of water irrigation systems, such as water storage, efficiency improvements, and systems to prevent loss.
Agriculture is currently the single largest use of water globally, Msangi explained, and it is likely to remain that way for the foreseeable future. As a result, water-related technologies are important to food production. Irrigation is the largest user of water, accounting for 70 percent of global water withdrawal and 90 percent of global water consumption (Shiklomanov, 2000). He said that agriculture, despite its large water use, is considered a residual claimant on water. In other words, the agricultural sector has access to water after other users. Higher value sectors tend to use water first, as they have a greater ability to pay for it. In the future, agriculture is likely to receive a smaller fraction of the water available, so the need for efficiency becomes more pressing.
Msangi posited that food security challenges are important and pressing as well. As the global population increases, income growth is also likely to increase in developing countries, causing an increased demand for high-value food (such as meat, fish, fruit, and vegetables). Msangi said that water resource distribution is uneven over aggregate regions of the world. Regions with large amounts of available water (such as Latin America) do not irrigate; they rely upon rain sources instead. Investors in irrigation technologies tend to be in the regions less endowed with water, such as South Asia, North Africa, and the Middle East.
Msangi then looked at irrigation water supply reliability under climate change predictions for 2050. He showed results from several different models, indicating that the ratio of supply to required water is decreasing because of both reduced availability and increased demand. The water reliability index is correspondingly decreasing. There is some variation in the model results, but in general, yield for crops of maize, rice, and wheat are all projected to decline in most regions of the world—in some cases, quite significantly. Growth in both income and population size drives food prices higher. Higher prices are the result of increased demand from the larger population, as well as the increased per-capita consumption that follows income growth. Msangi showed projections of crop price increases from 2010–2050 ranging from around 25 percent for wheat and rice to over 50 percent for maize (Nelson et al., 2010). Using
the mean effect from four different climate scenarios, those price increases double when adding the impact of climate change (Nelson et al., 2010).
Msangi then showed how variation to economic and population scenarios may impact crop prices. In an optimistic scenario (high income growth and low population growth), the price of rice increases the least, as Asian demand for rice falls with increasing income. In the most pessimistic scenario (low income growth and high population growth), the price of maize increases the most, as food demand rises in this socioeconomic scenario. In general, socioeconomic uncertainties contribute to price variations as much as the climate uncertainties. Critical uncertainties he pointed to include the following:
- Models used. There is significant variation in the outputs across models; climate outcomes are very different depending upon the model used. This includes region-specific variations in the effects.
- Socioeconomic and income growth. This growth is a key determinant for future food and water demand and price formation. Broad socioeconomic and well-being outcomes are also not clear. The best way to measure this is still under discussion, but higher food and water prices tend to lead to worse outcomes for the human population overall.
- Natural science and hydrology. An integrative approach linking the natural sciences with food and water sciences is needed.
- Land cover. Future land cover is uncertain, in terms of urbanization and livestock and land use change.
- Technology change. Technological improvements can determine productivity increases in food production and water efficiency.
Msangi stated it is important to better quantify the effects of climate change, particularly in its effect on the interrelationship between water and food supply.
James A. (Jae) Edmonds,
Chief Scientist and Battelle Laboratory Fellow,
Pacific Northwest National Laboratory
James A. Edmonds used an integrated assessment perspective to provide a view of what a 10-billion-person planet may look like in 2050. He began by showing the results from scenarios developed for the Intergovernmental Panel on Climate Change (IPCC) Sixth Assessment report, for
which working groups are beginning to prepare their findings. He said the scenario work was first used in the early 1990s to launch climate change assessments. They used representative concentration pathways (RCPs), which were landmark scenarios that were organized by the researchers. The second installment was organized by the community itself, which developed what are referred to as shared socioeconomic pathways (SSPs). Five SSP scenarios have been created designed to help understand the climate assessment problem (O’Neill et al., 2012):
- SSP1: Sustainability. It is easy to adapt to and mitigate climate change impact.
- SSP2: Middle of the road. Adaptation and mitigation are moderate challenges.
- SSP3: Fragmentation. It is difficult to adapt to and mitigate the effects of climate change. This is likely the world currently inhabited.
- SSP4: Inequality. It is difficult to adapt to but easy to mitigate climate change.
- SSP5: Conventional development. It is easy to adapt to but hard to mitigate challenges with climate change.
The SSPs are then used in an integrated assessment model, which uses information about population, labor, technology, and policies to provide quantified projections about prices, carbon emissions, energy supply, land use, livestock, and other economic and ecosystem variables. Edmonds explained the Global Change Assessment Model (GCAM) is technically complex and not optimized in any way—it gives results based purely upon the assumptions provided to it.
Edmonds then looked at the SSP assumptions for population, GDP, and technology. In terms of population, only SSP3 (fragmentation) actually has the population rise to 10 billion by 2050. The other models give rise to a population closer to 9 billion by that time. For GDP, the SSPs use International Development Association (IDA) labor productivity numbers to develop estimates. The GDP numbers are quite different based on the SSP chosen: SSP5 (conventional development) gives overall high GDP numbers, while SSP3 (fragmentation) yields the lowest GDP, and SSP4 (inequality) results in the greatest disparity in income. Regarding technology, the model includes assumptions about quality of life (such as energy demands and renewable technologies). Edmonds said that the technology assumptions also include information about crop yield improvement technologies; while this is a critical model assumption, it is not well studied. He indicated that land use policies also are important but not well studied.
Edmonds then showed some preliminary results for the different SSPs that assume no additional mitigation policies put into place. He showed radiative forcing (W/m2), a correlate of carbon dioxide emissions, for each SSP. SSP3 (fragmentation), the most likely current path, is on the high end of the scale; SSP5 (conventional development) results in a lower population but also in the highest radiative forcing, putting pressure on the climate system. The world of SSP5 is rich, with a lot of economic activity and a continued high use of fossil fuels. Edmonds showed the energy forms as a function of time for each scenario. There is a compositional change and variation across SSPs; however, the world is still dominated by fossil fuels in 2050 for all scenarios, particularly SSP5.
Looking at land use among the SSPs, Edmonds showed that there is little difference by percentage in land usage. Because urban land usage is such a small fraction relative to crops, forest, and biomass, it did not show up on his graphs.
Edmonds then introduced mitigation to the models. The mitigation information needs to specify the order in which countries implement the mitigation strategies; in general, wealthier nations start first, though some nations (Russia, Middle East) never begin mitigation implementation. Edmonds referred to this as the “delayed accession scenario.” The mitigation information also needs to specify whether there is policy in place for land use change; in other words, if there is no charge for indirect land use emissions (such as cutting down a tree). The introduction of mitigation plays out very differently in the five scenarios, which has implications for water demands.
Edmonds closed by stating that population can exert strong pressure on the Earth’s climate system, but economic activity, technology availability, and land use policies are also very important in determining the final impact on climate. In a world with 10 billion people, he stated, what matters is how much they are doing, how they are doing it, and what policies are in place to regulate activities.
The discussion period began with a question related to food consumption habits and the possible impact of changed food habits—for instance, increased demand for grass-fed beef. Msangi said the impact of livestock, particularly with respect to livestock intensively fed with grains, is not well understood. Grass-fed herds are not particularly well optimized; farmers in Brazil and Argentina say that they can increase herd densities, yet that is not included as a variable in any model. Reducing the need for feed grains, however, would take pressure off the production system. However, the stresses (such as heat, water, stocking density,
and food) on animal productivity and animal growth need to be better understood.
The discussion then moved to how rising food prices could affect inequality and the differences between urban and rural poor consumers. Msangi said his models do not take variations in the consumer into account. A more nuanced answer likely can come from the World Bank models, he suggested, which take into account different producer and consumer characteristics. In a more detailed case, milk producers in Latin America tend to benefit from higher food prices, though the benefit is not realized everywhere. In general, price increases are good for farmers in locations where land access is plentiful.
A participant expressed surprise that the price of rice is expected to rise only by 50 percent over the next 40 years. Msangi pointed out that this projected increase is in worldwide market prices, with greater price increases for domestic supply. The most striking price increase will occur in sub-Saharan Africa, and the effect is likely to spread to wheat as well.
A participant stated that current GDP growth is projected to be only four-fold by 2050, not the eight-fold increase referred to by Polasky. Polasky responded that he was not advocating for an eight-fold GDP increase, but rather provided a thought experiment to understand the numbers. He suggested turning the question around: Given an eight-fold increase in GDP cannot be attained, what should be done?
A participant asked about the importance of land grabbing, in which transnational governments, businesses, or individuals purchase land in large quantities, most commonly in developing countries (particularly in tropical regions). Edmonds said land grabbing is not typically assessed, but he does not believe it is an important consideration for the models, most of which look at the forces shaping land use, irrespective of ownership. Msangi pointed out that the model would need to include economic impacts; in that case, land grabbing may be more important because the profits are being expatriated. Land grabbing is a phenomenon documented by the World Bank and others, and the data are just now getting to the point where the phenomenon can be assessed analytically rather than anecdotally.
A participant brought up the topic of crop yield and productivity, asking about any easy ways to increase crop productivity, particularly in sub-Saharan Africa. Polasky responded that it is possible to increase food supply via intensification alone, but farmers have little incentive to do so. While not the case in the United States, he said, farmers are more likely to extend into undeveloped land than increase yield of their currently used land in Africa and tropical areas. Edmonds stated that there is a feedback effect against incentivizing increasing yield. Increasing the supply will not increase demand. Polasky also pointed to time lags in the
system; changes in the stock of natural capital (such as land use) take a while to have observable effects. For instance, looking at climate change, it is possible that enough emissions have been introduced to the point of irreversible damage, but the time lag in the system is so great that the negative impact will be felt only by future generations.