Phenotyping urban built and natural environments with high-resolution satellite images and unsupervised deep learning.

Cities in the developing world are expanding rapidly, and undergoing changes to their roads, buildings, vegetation, and other land use characteristics. Timely data are needed to ensure that urban change enhances health, wellbeing and sustainability. We present and evaluate a novel unsupervised deep clustering method to classify and characterise the complex and multidimensional built and natural environments of cities into interpretable clusters using high-resolution satellite images. We applied our approach to a high-resolution (0.3 m/pixel) satellite image of Accra, Ghana, one of the fastest growing cities in sub-Saharan Africa, and contextualised the results with demographic and environmental data that were not used for clustering. We show that clusters obtained solely from images capture distinct interpretable phenotypes of the urban natural (vegetation and water) and built (building count, size, density, and orientation; length and arrangement of roads) environment, and population, either as a unique defining characteristic (e.g., bodies of water or dense vegetation) or in combination (e.g., buildings surrounded by vegetation or sparsely populated areas intermixed with roads). Clusters that were based on a single defining characteristic were robust to the spatial scale of analysis and the choice of cluster number, whereas those based on a combination of characteristics changed based on scale and number of clusters. The results demonstrate that satellite data and unsupervised deep learning provide a cost-effective, interpretable and scalable approach for real-time tracking of sustainable urban development, especially where traditional environmental and demographic data are limited and infrequent.

Quantifying within-city inequalities in child mortality across neighbourhoods in Accra, Ghana: a Bayesian spatial analysis.

OBJECTIVE: Countries in sub-Saharan Africa suffer the highest rates of child mortality worldwide. Urban areas tend to have lower mortality than rural areas, but these comparisons likely mask large within-city inequalities. We aimed to estimate rates of under-five mortality (U5M) at the neighbourhood level for Ghana's Greater Accra Metropolitan Area (GAMA) and measure the extent of intraurban inequalities. METHODS: We accessed data on >700 000 women aged 25-49 years living in GAMA using the most recent Ghana census (2010). We summarised counts of child births and deaths by five-year age group of women and neighbourhood (n=406) and applied indirect demographic methods to convert the summaries to yearly probabilities of death before age five years. We fitted a Bayesian spatiotemporal model to the neighbourhood U5M probabilities to obtain estimates for the year 2010 and examined their correlations with indicators of neighbourhood living and socioeconomic conditions. RESULTS: U5M varied almost five-fold across neighbourhoods in GAMA in 2010, ranging from 28 (95% credible interval (CrI) 8 to 63) to 138 (95% CrI 111 to 167) deaths per 1000 live births. U5M was highest in neighbourhoods of the central urban core and industrial areas, with an average of 95 deaths per 1000 live births across these neighbourhoods. Peri-urban neighbourhoods performed better, on average, but rates varied more across neighbourhoods compared with neighbourhoods in the central urban areas. U5M was negatively correlated with multiple indicators of improved living and socioeconomic conditions among peri-urban neighbourhoods. Among urban neighbourhoods, correlations with these factors were weaker or, in some cases, reversed, including with median household consumption and women's schooling. CONCLUSION: Reducing child mortality in high-burden urban neighbourhoods in GAMA, where a substantial portion of the urban population resides, should be prioritised as part of continued efforts to meet the Sustainable Development Goal national target of less than 25 deaths per 1000 live births.

Synthesizing the links between secure housing tenure and health for more equitable cities.

Millions of households in rich and poor countries alike are at risk of being unwilfully displaced from their homes or the land on which they live (i.e., lack secure tenure), and the urban poor are most vulnerable. Improving housing tenure security may be an intervention to improve housing and environmental conditions and reduce urban health inequalities. Building on stakeholder workshops and a narrative review of the literature, we developed a conceptual model that infers the mechanisms through which more secure housing tenure can improve housing, environmental quality, and health. Empirical studies show that more secure urban housing tenure can boost economic mobility, improve housing and environmental conditions including reduced exposure to pollution, create safer and more resourced communities, and improve physical and mental health. These links are shared across tenure renters and owners and different economic settings. Broader support is needed for context-appropriate policies and actions to improve tenure security as a catalyst for cultivating healthier homes and neighbourhoods and reducing urban health inequalities in cities.

Time and space catch up with restoration programs that ignore ecosystem service trade-offs.

In response to extreme societal consequences of ecosystem degradation and climate change, attention to ecological restoration is increasing globally. In China, investments in restoration exceeded USD 378.5 billion over the past decade. However, restoration programs are experiments that can cause marked unintended consequences, with trade-offs across space and time that have undergone little empirical examination. We quantified the long-term effects of large-scale afforestation for soil erosion and sandstorm prevention in semiarid China. We found that soil erosion was notably reduced by afforestation but surface runoff declined significantly, after a time lag of 18 years, limiting overall benefit. While forest area also increased, forest quality declined, interacting with reduced surface water runoff. Crucially, increased forest water consumption accelerated downstream groundwater depletion, thus intensifying conflicts over water use. The time lags and spatial trade-offs revealed by this case study provide critical lessons for large-scale restoration programs globally.

Participatory Modeling of Water Vulnerability in Remote Alaskan Households Using Causal Loop Diagrams.

Despite perceptions of high water availability, adequate access to sufficient water resources remains a major challenge in Alaska. This paper uses a participatory modeling approach to investigate household water vulnerability in remote Alaska and to examine factors that affect water availability and water access. Specifically, the work asks: how do water policy stakeholders conceptualize the key processes that affect household water vulnerability in the context of rural Alaska? Fourteen water policy stakeholders participated in the modeling process, which included defining the problem of household water vulnerability and constructing individual causal loop diagrams (CLDs) that represent their conceptualization of household water vulnerability. Individual CLDs were subsequently combined and five sub-models emerged: environmental, economic, infrastructure, social, and health. The environmental and economic sub-models of the CLD are explored in depth. In the environmental sub-model, climate change and environmental barriers due to geography influence household water vulnerability. In the economic sub-model, four processes and one feedback loop affect household water vulnerability, including operations and maintenance funding, the strength of the rural Alaskan economy, and the impact of regulations. To overcome household water vulnerability and make households more resilient, stakeholders highlighted policy solutions under five themes: economics, social, regulatory, technological, and environmental.

Realizing the values of natural capital for inclusive, sustainable development: Informing China's new ecological development strategy.

A major challenge in transforming development to inclusive, sustainable pathways is the pervasive and persistent trade-off between provisioning services (e.g., agricultural production) on the one hand and regulating services (e.g., water purification, flood control) and biodiversity conservation on the other. We report on an application of China's new Ecological Development Strategy, now being formally tested and refined for subsequent scaling nationwide, which aims to mitigate and even eliminate these trade-offs. Our focus is the Ecosystem Function Conservation Area of Hainan Island, a rural, tropical region where expansion of rubber plantations has driven extensive loss of natural forest and its vital benefits to people. We explored both the biophysical and the socioeconomic options for achieving simultaneous improvements in product provision, regulating services, biodiversity, and livelihoods. We quantified historic trade-offs between rubber production and vital regulating services, finding that, over the past 20 y (1998-2017), there was a 72.2% increase in rubber plantation area, leading to decreases in soil retention (17.8%), water purification [reduced retention of nitrogen (56.3%) and phosphorus (27.4%)], flood mitigation (21.9%), carbon sequestration (1.7%), and habitat for biodiversity (6.9%). Using scenario analyses, we identified a two-pronged strategy that would significantly reduce these trade-offs, enhancing regulating services and biodiversity, while simultaneously diversifying and increasing product provision and improving livelihoods. This general approach to analyzing product provision, regulating services, biodiversity, and livelihoods has applicability in rural landscapes across China, South and Southeast Asia, and beyond.

Reimagining the potential of Earth observations for ecosystem service assessments.

The benefits nature provides to people, called ecosystem services, are increasingly recognized and accounted for in assessments of infrastructure development, agricultural management, conservation prioritization, and sustainable sourcing. These assessments are often limited by data, however, a gap with tremendous potential to be filled through Earth observations (EO), which produce a variety of data across spatial and temporal extents and resolutions. Despite widespread recognition of this potential, in practice few ecosystem service studies use EO. Here, we identify challenges and opportunities to using EO in ecosystem service modeling and assessment. Some challenges are technical, related to data awareness, processing, and access. These challenges require systematic investment in model platforms and data management. Other challenges are more conceptual but still systemic; they are byproducts of the structure of existing ecosystem service models and addressing them requires scientific investment in solutions and tools applicable to a wide range of models and approaches. We also highlight new ways in which EO can be leveraged for ecosystem service assessments, identifying promising new areas of research. More widespread use of EO for ecosystem service assessment will only be achieved if all of these types of challenges are addressed. This will require non-traditional funding and partnering opportunities from private and public agencies to promote data exploration, sharing, and archiving. Investing in this integration will be reflected in better and more accurate ecosystem service assessments worldwide.

Soil-targeted interventions could alleviate locust and grasshopper pest pressure in West Africa.

Agricultural land use has intended and unintended consequences for human livelihoods through feedbacks within coupled human and natural systems. In Senegal, West Africa, soils are a vital resource for livelihoods and food security in smallholder farming communities. In this study, we explored the connections among land use, soil conditions, plant nutrient content, and the abundance of several locust and grasshopper species. We worked in two rural farming villages in the Kaffrine region of Senegal. Oedaleus senegalensis was least abundant in groundnut areas where plant N was highest and abundance was negatively correlated with plant N across land use types. Overall, grasshoppers were most numerous in grazing and fallow areas. There was little variation in soil properties across land use types and soil organic matter (SOM) and inorganic soil N content were low throughout. SOM was positively correlated with soil inorganic N concentration, which in turn was positively correlated with plant N content. Of the management practices we surveyed, fallowing fields was important for soil N and SOM replenishment. These results corroborate other research indicating that land use, management practices, soil and plant nutrients, and insect herbivore abundance are mechanistically coupled. Although further research is needed, improving soil fertility could be used as an alternative to pesticides to keep locusts at bay and improve crop yields.

Nitrogen balance dynamics during 2000-2010 in the Yangtze River Basin croplands, with special reference to the relative contributions of cropland area and synthetic fertilizer N application rate changes.

With the increases of cropland area and fertilizer nitrogen (N) application rate, general N balance characteristics in regional agroecosystems have been widely documented. However, few studies have quantitatively analyzed the drivers of spatial changes in the N budget. We constructed a mass balance model of the N budget at the soil surface using a database of county-level agricultural statistics to analyze N input, output, and proportional contribution of various factors to the overall N input changes in croplands during 2000-2010 in the Yangtze River Basin, the largest basin and the main agricultural production region in China. Over the period investigated, N input increased by 9%. Of this 87% was from fertilizer N input. In the upper and middle reaches of the basin, the increased synthetic fertilizer N application rate accounted for 84% and 76% of the N input increase, respectively, mainly due to increased N input in the cropland that previously had low synthetic fertilizer N application rate. In lower reaches of the basin, mainly due to urbanization, the decrease in cropland area and synthetic fertilizer N application rate nearly equally contributed to decreases in N input. Quantifying spatial N inputs can provide critical managerial information needed to optimize synthetic fertilizer N application rate and monitor the impacts of urbanization on agricultural production, helping to decrease agricultural environment risk and maintain sustainable agricultural production in different areas.

Conservation vs. livelihoods: spatial management of non-timber forest product harvests in a two-dimensional model.

Areas of high biodiversity often coincide with communities living in extreme poverty. As a livelihood support, these communities often harvest wild products from the environment. But harvest activities can have negative impacts on fragile and globally important ecosystems. This paper examines trade-offs in ecological protection and community welfare from the harvest of wild products. With a novel model and empirical evidence, I show that management of harvest activity does not always resolve these trade-offs. In a model of continuous harvests in a two-dimensional landscape, managed harvest activity improves welfare, but is uniformly bad for other ecosystem services that are sensitive to the presence (as opposed to the intensity) of human activity. Empirical results from a unique dataset of mushroom harvesters in Yunnan, China suggest more experienced, poorer, and more vulnerable individuals tend to rely on more distant harvests. Thus, policies that limit the extent of forest travel, such as protected areas, may protect fragile ecosystems but can have a disproportionately negative effect on those most vulnerable.

Meta-studies in land use science: Current coverage and prospects.

Land use science has traditionally used case-study approaches for in-depth investigation of land use change processes and impacts. Meta-studies synthesize findings across case-study evidence to identify general patterns. In this paper, we provide a review of meta-studies in land use science. Various meta-studies have been conducted, which synthesize deforestation and agricultural land use change processes, while other important changes, such as urbanization, wetland conversion, and grassland dynamics have hardly been addressed. Meta-studies of land use change impacts focus mostly on biodiversity and biogeochemical cycles, while meta-studies of socioeconomic consequences are rare. Land use change processes and land use change impacts are generally addressed in isolation, while only few studies considered trajectories of drivers through changes to their impacts and their potential feedbacks. We provide a conceptual framework for linking meta-studies of land use change processes and impacts for the analysis of coupled human-environmental systems. Moreover, we provide suggestions for combining meta-studies of different land use change processes to develop a more integrated theory of land use change, and for combining meta-studies of land use change impacts to identify tradeoffs between different impacts. Land use science can benefit from an improved conceptualization of land use change processes and their impacts, and from new methods that combine meta-study findings to advance our understanding of human-environmental systems.

Benefits, costs, and livelihood implications of a regional payment for ecosystem service program.

Despite broad interest in using payment for ecosystem services to promote changes in the use of natural capital, there are few expost assessments of impacts of payment for ecosystem services programs on ecosystem service provision, program cost, and changes in livelihoods resulting from program participation. In this paper, we evaluate the Paddy Land-to-Dry Land (PLDL) program in Beijing, China, and associated changes in service providers' livelihood activities. The PLDL is a land use conversion program that aims to protect water quality and quantity for the only surface water reservoir that serves Beijing, China's capital city with nearly 20 million residents. Our analysis integrates hydrologic data with household survey data and shows that the PLDL generates benefits of improved water quantity and quality that exceed the costs of reduced agricultural output. The PLDL has an overall benefit-cost ratio of 1.5, and both downstream beneficiaries and upstream providers gain from the program. Household data show that changes in livelihood activities may offset some of the desired effects of the program through increased expenditures on agricultural fertilizers. Overall, however, reductions in fertilizer leaching from land use change dominate so that the program still has a positive net impact on water quality. This program is a successful example of water users paying upstream landholders to improve water quantity and quality through land use change. Program evaluation also highlights the importance of considering behavioral changes by program participants.