The rapid growth of remotely sensed earth observation data presents clear opportunities for monitoring complex ecosystem change and answering fundamental ecological questions. However, large-scale automated monitoring of ecosystems faces challenges. Data-driven models require extensive datasets and often lack generalizability when training data are unrepresentative, while process-driven models, such as radiative transfer models (RTMs), can be imprecise due to gaps in knowledge, or simplified representation of physical processes. To enhance the prediction of plant functional traits and simultaneously discover where process-driven models can be improved, we explore the potential of Physics-Informed Neural Networks (PINNs) as a hybrid approach that combines the strengths of both methodologies at the leaf scale. In contrast to data augmentation approaches, our implementation directly integrates the widely-used PROSPECT5B model into the architecture of an autoencoder framework. Our results show that our PINNs approach is able to outperform data-driven techniques even when trained on very limited training data (i.e. 17 % training vs 83 % validation). We also identified weak points in the PROSPECT5B model by progressively replacing individual components of PROSPECT5B with convolutional neural networks. Our case study indicates that especially Prospect's generalized “plate model” could be refined to improve predictive ability. Hence, our framework provides a self-diagnostic capability and identifies areas for improvement in process-driven models and their components. Thus, we conclude that PINNs 1) improve data-driven predictive accuracy while maintaining physical consistency with minimal training data while 2) being able to identify limitations in process-driven models. Hence, we believe our framework could serve as a new standard for evolving and improving radiative transfer models.

Ecosystems are threatened by increasing droughts under climate change. A multitude of plant physiological regulation processes determine the overall drought resistance of ecosystems. So far, these physiological strategies to resist drought are poorly understood at large scales across different ecosystem types because the detection of these physiological regulation processes is mostly limited to in situ measurements on individual plants. In this study, by using high-resolution remote sensing data, we evaluated drought strategies of different ecosystem types throughout Europe by evaluating three key physiological regulation aspects (evapotranspiration, water content, and carbon regulation) based on their associated vegetation attributes. We found that different ecosystem types show divergent responses in these physiological attributes, suggesting different optimization strategies with respect to water saving versus spending, water content stabilizing versus fluctuating, and leaf conserving versus shedding strategies facing drought. These drought strategies from remote sensing provide timely ecosystem response information, facilitating earth system model predictions and aiding the protection against future droughts at large scales.

Increased climate impacts threaten coastal functions globally, highlighting the need for multifunctional coastal climate adaptation. Sand nourishment can adapt sandy coasts to sea level rise, mitigate erosion, increase flood safety, enhance ecological habitats and expand recreational space. Therefore, sand nourishment is increasingly regarded as a promising nature-based strategy for coastal climate adaptation. However, despite this growing recognition, the assessment of how sand nourishment design impacts multifunctional adaptation remains limited. In this perspective article, we argue for three key lessons for researchers to optimise assessing multifunctional coastal climate adaptation by sand nourishment. We conducted stakeholder workshops to scope and inform our perspective, performed semi-structured literature reviews to concretise and validate this for international applications, built a qualitative model to visualise our interdisciplinary overview of how nourishments impact coastal multifunctionality, reflected on this in expert workshops, and identified implications for researchers. In this manner, we assessed the effects of nourishment design on coastal morphology, ecology, socio-economics and ecosystem services in realising the key policy goals of flood safety, nature and recreation. We found that sand nourishment design can result in conflicts between policy goals, generate ambiguous outcomes and lead to system-wide feedback effects. As such, we identified three key lessons: (1) conflicts between policy goals require informing political decision-making on prioritisation between coastal functions, (2) concreteness is needed on otherwise ambiguous functions, and (3) ongoing, multidisciplinary system-wide monitoring is essential. We thus call for a holistic approach to sand nourishment design and encourage researchers from diverse expertise and localities to expand on and adapt our findings to optimise informing sand nourishment design for delivering multifunctional coastal climate adaptation worldwide.

While wild pollinators play a key role in global food production, their assessment is currently missing from the most commonly used environmental impact assessment method, Life Cycle Assessment (LCA). This is mainly due to constraints in data availability and compatibility with LCA inventories. To target this gap, relative pollinator abundance estimates were obtained with the use of a Delphi assessment, during which 25 experts, covering 16 nationalities and 45 countries of expertise, provided scores for low, typical, and high expected abundance associated with 24 land use categories. Based on these estimates, this study presents a set of globally generic characterization factors (CFs) that allows translating land use into relative impacts to wild pollinator abundance. The associated uncertainty of the CFs is presented along with an illustrative case to demonstrate the applicability in LCA studies. The CFs based on estimates that reached consensus during the Delphi assessment are recommended as readily applicable and allow key differences among land use types to be distinguished. The resulting CFs are proposed as the first step for incorporating pollinator impacts in LCA studies, exemplifying the use of expert elicitation methods as a useful tool to fill data gaps that constrain the characterization of key environmental impacts.

Green infrastructure (GI) is increasingly addressed in urban planning and research to enhance urban sustainability and resilience through the provisioning of ecosystem services (ES). Yet, few applications exist of planning models for multifunctional GI in high spatial and thematic detail that simultaneously align with stakeholder interests. We address these gaps by developing and presenting a spatially explicit model to inform urban planners on priority areas for multifunctional GI development. This model was made possible by spatial analyses on multiple scales, enabling us to assess ES in sufficient detail, while simultaneously matching the preferences for scale and ES-indicators of decision makers and urban planners. The model involves a novel weighting scheme based upon the local capacity of GI to mitigate problems. We applied our model to the city of The Hague using a set of three policy-relevant problems: air pollution, the urban heat island effect and storm water flooding. Our results show that the capacity of GI to mitigate these problems varies spatially, both within and between ES, and depends on local characteristics of GI and the environmental context. We illustrate the relevance of using a multi-scale approach in spatial ES analysis, and underline that GI planning measures should be assessed in high spatial detail due to their often locally distinct ES capacity. Our approach makes important strides towards the deployment of nature-based solutions for urban challenges in the light of demands for increasing resilience and sustainability.

Ecosystem services (ES) are increasingly embedded in policy agendas, but if and how policy actors are considering them is not often reported. This study assesses the extent to which ES were considered by key policy actors involved in the strategic decision-making process leading to an innovative large-scale Dutch coastal management project. We analysed retrospective interviews to ascertain which ES were considered and how they were described by policy actors. Over half of the quotes (118/228) and 16 out of the 17 interviewees referred to three broad ES categories, with high degrees of adoption: coastal safety, recreation and cognitive development (learning by doing). The broad terms ‘nature’ and ‘spatial quality’ were also referenced often (36 times). Our findings suggest that broad, unspecified ecosystem services were adopted highly by the policy actors, while specific ecosystem service categories were rarely considered. Relatable and comprehensible cultural ecosystem services also constituted critical arguments for policy actors in their strategic decision making. We reflect that ambiguous, broad terms can help to garner support and unite efforts across disciplinary and institutional boundaries. For ES to align with relevant aspects of decision making, a ‘translation step’ between ES research and decision making might be required and ambiguity should be acknowledged.