Kathleen C. Weathers

Wildland fire smoke and ash are pervasive and a significant source of particulate matter to the atmosphere. In addition to effects on air quality, wildland smoke and ash deposit to ecosystems via dry (particulate) and wet (precipitation) deposition. Here we combine remotely sensed smoke polygon data with rainfall and weekly wet deposition data from > 250 U.S. National Atmospheric Deposition Program (NADP) sites to quantify wildland smoke influence on wet ion deposition during 2014, 2020, and 2022. Combined records of daily smoke and rain indicate that sites experienced 21 smoke‐rain days (when smoke and rainfall co‐occur), on average, in 2022 compared to 4 days in 2014 and 6 days in 2020. The number of smoke‐rain days and their spatial distribution were driven more by variation in smoke than rain occurrence. Across 16 climate regions, the Upper Midwest was a persistent hotspot for smoke‐rain days in all 3 years. Mixed effects models estimating site‐level impacts of smoke‐rain on a weekly basis showed that smoke‐rain days had significant positive effects on wet deposition of all nine ions evaluated. Orthophosphate (PO ₄ ³⁻ , 46%), potassium (K ⁺ , 41%), and ammonium (NH ₄ ⁺ , 36%) deposition were most enhanced compared to non‐smoke‐rain days. The magnitude of these effects varied by region. With each additional smoke‐rain day, weekly wet NH ₄ ⁺ deposition increased by 100%–220% in Alaska & the Arctic, Northwest, and West regions, whereas PO ₄ ³⁻ deposition increased ~100% in the Northeast. Based on modeled estimates, smoke‐rain days contributed ~20%–30% of annual wet NH ₄ ⁺ , PO ₄ ³⁻ , K ⁺ , and Ca ²⁺ deposition in 2022. Further, molar N:P ratios of wet deposition were higher at 76% of NADP sites during weeks with versus without smoke‐rain. Our results demonstrate that smoke‐affected wet deposition can deliver biologically important nutrients to ecosystems, which may be leading to fire‐fertilizing of terrestrial and aquatic ecosystems.

Given the critical importance of freshwater for human well‐being and the myriad threats altering water quality and quantity, monitoring lakes and reservoirs is necessary to identify where, how, and why freshwater ecosystems are changing globally. Satellite remote sensing tools hold great promise for improving water quality monitoring to advance research and management objectives, but “ready‐to‐use” solutions are not yet available for navigating the decisions required to collect, process, and analyze quality remote sensing data. While new products and techniques are continuously being developed, their use for research and management remains challenging. Aspiring users of satellite data face key decisions among many different options to select, process, and use these data to estimate surface water quality parameters. These decisions have important downstream consequences for inference, but knowledge gaps remain in how to navigate them to match project goals, particularly for interdisciplinary teams. We point to key considerations for how users can evaluate the spatiotemporal requirements for their specific project and identify essential decision points in how to process and interpret the acquired satellite data to estimate water quality parameters, using cyanobacterial blooms as a focal example. Well‐considered front‐end decisions are crucial for matching satellite and in situ data and fully leveraging satellite data to address the specific question and system. Importantly, the science and tools underlying remote sensing methods are evolving quickly and the “best” choices depend on the study or monitoring objective. We provide information for nonspecialists and questions for interdisciplinary teams to consider to aid in the wise use of this transformative technology. We conclude by identifying five priorities for further development and emphasizing the importance of reciprocal knowledge exchange and engagement between the aquatic ecology and remote sensing communities.

Studies of lake ice phenology have historically relied on limited in situ data. Relatively few observations exist for ice out and fewer still for ice in, both of which are necessary to determine the temporal extent of ice cover. Satellite data provide an opportunity to better document patterns of ice phenology across landscapes and relate them to the climatological drivers behind changing ice phenology. We developed a model, the Cumulative Sum Method (CSM), that uses daytime and nighttime surface temperature observations from the Moderate Resolution Imaging Spectroradiometer (MODIS) sensor on board the Earth‐observing Aqua satellite to approximate ice in (the onset of ice cover) and ice out from training datasets of 13 and 58 Maine lakes, respectively, during the 2002/2003 through 2017/2018 ice seasons. Ice in was signaled by reaching a threshold of cumulative negative degrees following the first day of the season below 0°C. Ice out was signaled by reaching a threshold of cumulative positive degrees following the first day of the year above 0°C. The comparison of observed and remotely sensed ice‐in dates showed relative agreement with a correlation coefficient of 0.71 and a mean absolute error (MAE) of 9.8 days. Ice‐out approximations had a correlation coefficient of 0.67 and an MAE of 8.8 days. Lakes smaller in surface area and nearer the Atlantic coast had the greatest error in approximation. Application of the CSM to 20 additional lakes in Maine produced a comparable ice‐out MAE of 8.9 days. Ice‐out model performance was weaker for the warmest years; there was a larger MAE of 12.0 days when the model was applied to the years 2019–2023 for the original 58 lakes. The development of this model, which utilizes daily satellite data, demonstrates the promise of remote sensing for quantifying ice phenology over short, temporal scales, and wider geographic regions than can be observed in situ, and allows exploration of the influence of surface temperature patterns on the process and timing of ice in and ice out.

Research in both ecology and AI strives for predictive understanding of complex systems, where nonlinearities arise from multidimensional interactions and feedbacks across multiple scales. After a century of independent, asynchronous advances in computational and ecological research, we foresee a critical need for intentional synergy to meet current societal challenges against the backdrop of global change. These challenges include understanding the unpredictability of systems-level phenomena and resilience dynamics on a rapidly changing planet. Here, we spotlight both the promise and the urgency of a convergence research paradigm between ecology and AI. Ecological systems are a challenge to fully and holistically model, even using the most prominent AI technique today: deep neural networks. Moreover, ecological systems have emergent and resilient behaviors that may inspire new, robust AI architectures and methodologies. We share examples of how challenges in ecological systems modeling would benefit from advances in AI techniques that are themselves inspired by the systems they seek to model. Both fields have inspired each other, albeit indirectly, in an evolution toward this convergence. We emphasize the need for more purposeful synergy to accelerate the understanding of ecological resilience whilst building the resilience currently lacking in modern AI systems, which have been shown to fail at times because of poor generalization in different contexts. Persistent epistemic barriers would benefit from attention in both disciplines. The implications of a successful convergence go beyond advancing ecological disciplines or achieving an artificial general intelligence—they are critical for both persisting and thriving in an uncertain future.

Although remote sensing of lake surface water temperature has been ongoing for decades and has led to important discoveries regarding the warming of lake surface temperature in some regions, the ability to access and use remote sensing data is still primarily limited to remote sensing experts. Effective use of remote sensing data involves technical skills in coding (often in multiple programming languages), application of appropriate atmospheric corrections, and integrating spatially heterogeneous remote sensing data with in situ data obtained at specific geographic locations. To improve access to remote sensing data and broaden the understanding of changes in lake surface water temperature over the past four decades, we created lakeCoSTR (lake Collection 2 Surface Temperature Retrieval), a user‐friendly, cloud‐based script that gives ecologists lacking specialized training in remote sensing the ability to access the Landsat Collection 2 temperature estimates for lakes with a surface area of at least 0.4 ha. Additionally, if in situ data are provided, a paired dataset can be created within the tool. To demonstrate lakeCoSTR, we retrieved surface water temperature data for a lake with a long monitoring history, Lake Sunapee, NH, USA, and compared long‐term surface temperature trends between data obtained via lakeCoSTR and in situ measurements. When compared with Landsat Collection 1 temperature estimates derived from a single‐channel algorithm, the Landsat Collection 2 data from lakeCoSTR required no calibration to in situ data. This suggests that lakeCoSTR can be used to document temporal trends in lakes and to reprocess analyses that relied on Collection 1 data. From 1983 until 2020, remotely sensed surface water temperature estimates at Lake Sunapee increased by 0.07–0.09°C year⁻¹ during July and August, a trend that agreed with long‐term in situ record. As climate change continues to impact freshwater systems, a better understanding of long‐term temperature data will be vital. Remotely sensed data, like those acquired using lakeCoSTR, may provide a window into temperature trends for lakes.

The concentration of dissolved organic nitrogen (DON) is increasing in many northern hemisphere lakes, yet its use by phytoplankton and fate in the environment seldom have been quantified. We conducted 1 week, in situ, microcosm incubations across 25 lakes in northeastern North America to understand how DON, dissolved inorganic nitrogen (DIN), and dissolved inorganic phosphorus (P) affected phytoplankton biomass. In addition, we tested whether lakes were limited by single macronutrients (N or P) or colimited by both. Phytoplankton biomass in 80% of lakes responded similarly to DON and DIN additions. Of the lakes where N form produced differential responses, the majority of phytoplankton communities exhibited greater biomass accumulation with DON than DIN. Colimitation was the most common type of nutrient limitation among the study lakes, followed by P limitation. Limitation type shifted with N form in 40% of the study lakes, but without consistent patterns explaining how shifts occurred. Regardless of N form, lakes with watersheds more dominated by agriculture and higher total dissolved nitrogen (TDN) tended to show P‐limited phytoplankton responses, while lakes with less agricultural watersheds and lower TDN tended to show colimited phytoplankton responses. Finally, ambient TDN and total phosphorus (TP) nutrient concentrations were stronger predictors of limitation type than ambient TDN : TP ratios. The different contributions of DON and DIN to phytoplankton biomass in some of our study lakes suggest that DON loading from surrounding watersheds may be an overlooked component in predicting phytoplankton productivity and nutrient limitation dynamics in freshwater ecosystems.

Freshwater ecosystems provide vital services, yet are facing increasing risks from global change. In particular, lake thermal dynamics have been altered around the world as a result of climate change, necessitating a predictive understanding of how climate will continue to alter lakes in the future as well as the associated uncertainty in these predictions. Numerous sources of uncertainty affect projections of future lake conditions but few are quantified, limiting the use of lake modeling projections as management tools. To quantify and evaluate the effects of two potentially important sources of uncertainty, lake model selection uncertainty and climate model selection uncertainty, we developed ensemble projections of lake thermal dynamics for a dimictic lake in New Hampshire, USA (Lake Sunapee). Our ensemble projections used four different climate models as inputs to five vertical one-dimensional (1-D) hydrodynamic lake models under three different climate change scenarios to simulate thermal metrics from 2006 to 2099. We found that almost all the lake thermal metrics modeled (surface water temperature, bottom water temperature, Schmidt stability, stratification duration, and ice cover, but not thermocline depth) are projected to change over the next century. Importantly, we found that the dominant source of uncertainty varied among the thermal metrics, as thermal metrics associated with the surface waters (surface water temperature, total ice duration) were driven primarily by climate model selection uncertainty, while metrics associated with deeper depths (bottom water temperature, stratification duration) were dominated by lake model selection uncertainty. Consequently, our results indicate that researchers generating projections of lake bottom water metrics should prioritize including multiple lake models for best capturing projection uncertainty, while those focusing on lake surface metrics should prioritize including multiple climate models. Overall, our ensemble modeling study reveals important information on how climate change will affect lake thermal properties, and also provides some of the first analyses on how climate model selection uncertainty and lake model selection uncertainty interact to affect projections of future lake dynamics.

Freshwater ecosystems provide vital services, yet are facing increasing risks from global change. In particular, lake thermal dynamics have been altered around the world as a result of climate change, necessitating a predictive understanding of how climate will continue to alter lakes in the future as well as the associated uncertainty in these predictions. Numerous sources of uncertainty affect projections of future lake conditions but few are quantified, limiting the use of lake modeling projections as management tools. To quantify and evaluate the effects of two potentially important sources of uncertainty, lake model selection uncertainty and climate model selection uncertainty, we developed ensemble projections of lake thermal dynamics for a dimictic lake in New Hampshire, USA (Lake Sunapee). Our ensemble projections used four different climate models as inputs to five vertical one-dimensional (1-D) hydrodynamic lake models under three different climate change scenarios to simulate thermal metrics from 2006 to 2099. We found that almost all the lake thermal metrics modeled (surface water temperature, bottom water temperature, Schmidt stability, stratification duration, and ice cover, but not thermocline depth) are projected to change over the next century. Importantly, we found that the dominant source of uncertainty varied among the thermal metrics, as thermal metrics associated with the surface waters (surface water temperature, total ice duration) were driven primarily by climate model selection uncertainty, while metrics associated with deeper depths (bottom water temperature, stratification duration) were dominated by lake model selection uncertainty. Consequently, our results indicate that researchers generating projections of lake bottom water metrics should prioritize including multiple lake models for best capturing projection uncertainty, while those focusing on lake surface metrics should prioritize including multiple climate models. Overall, our ensemble modeling study reveals important information on how climate change will affect lake thermal properties, and also provides some of the first analyses on how climate model selection uncertainty and lake model selection uncertainty interact to affect projections of future lake dynamics.