Qualitative crop condition survey reveals spatiotemporal production patterns and allows early yield prediction.

Large-scale continuous crop monitoring systems (CMS) are key to detect and manage agricultural production anomalies. Current CMS exploit meteorological and crop growth models, and satellite imagery, but have underutilized legacy sources of information such as operational crop expert surveys with long and uninterrupted records. We argue that crop expert assessments, despite their subjective and categorical nature, capture the complexities of assessing the "status" of a crop better than any model or remote sensing retrieval. This is because crop rating data naturally encapsulates the broad expert knowledge of many individual surveyors spread throughout the country, constituting a sophisticated network of "people as sensors" that provide consistent and accurate information on crop progress. We analyze data from the US Department of Agriculture (USDA) Crop Progress and Condition (CPC) survey between 1987 and 2019 for four major crops across the US, and show how to transform the original qualitative data into a continuous, probabilistic variable better suited to quantitative analysis. Although the CPC reflects the subjective perception of many surveyors at different locations, the underlying models that describe the reported crop status are statistically robust and maintain similar characteristics across different crops, exhibit long-term stability, and have nation-wide validity. We discuss the origin and interpretation of existing spatial and temporal biases in the survey data. Finally, we propose a quantitative Crop Condition Index based on the CPC survey and demonstrate how this index can be used to monitor crop status and provide earlier and more precise predictions of crop yields than official USDA forecasts released midseason.

Wildfires and climate change push low-elevation forests across a critical climate threshold for tree regeneration.

Climate change is increasing fire activity in the western United States, which has the potential to accelerate climate-induced shifts in vegetation communities. Wildfire can catalyze vegetation change by killing adult trees that could otherwise persist in climate conditions no longer suitable for seedling establishment and survival. Recently documented declines in postfire conifer recruitment in the western United States may be an example of this phenomenon. However, the role of annual climate variation and its interaction with long-term climate trends in driving these changes is poorly resolved. Here we examine the relationship between annual climate and postfire tree regeneration of two dominant, low-elevation conifers (ponderosa pine and Douglas-fir) using annually resolved establishment dates from 2,935 destructively sampled trees from 33 wildfires across four regions in the western United States. We show that regeneration had a nonlinear response to annual climate conditions, with distinct thresholds for recruitment based on vapor pressure deficit, soil moisture, and maximum surface temperature. At dry sites across our study region, seasonal to annual climate conditions over the past 20 years have crossed these thresholds, such that conditions have become increasingly unsuitable for regeneration. High fire severity and low seed availability further reduced the probability of postfire regeneration. Together, our results demonstrate that climate change combined with high severity fire is leading to increasingly fewer opportunities for seedlings to establish after wildfires and may lead to ecosystem transitions in low-elevation ponderosa pine and Douglas-fir forests across the western United States.

Multispectral high resolution sensor fusion for smoothing and gap-filling in the cloud.

Remote sensing optical sensors onboard operational satellites cannot have high spectral, spatial and temporal resolutions simultaneously. In addition, clouds and aerosols can adversely affect the signal contaminating the land surface observations. We present a HIghly Scalable Temporal Adaptive Reflectance Fusion Model (HISTARFM) algorithm to combine multispectral images of different sensors to reduce noise and produce monthly gap free high resolution (30 m) observations over land. Our approach uses images from the Landsat (30 m spatial resolution and 16 day revisit cycle) and the MODIS missions, both from Terra and Aqua platforms (500 m spatial resolution and daily revisit cycle). We implement a bias-aware Kalman filter method in the Google Earth Engine (GEE) platform to obtain fused images at the Landsat spatial-resolution. The added bias correction in the Kalman filter estimates accounts for the fact that both model and observation errors are temporally auto-correlated and may have a non-zero mean. This approach also enables reliable estimation of the uncertainty associated with the final reflectance estimates, allowing for error propagation analyses in higher level remote sensing products. Quantitative and qualitative evaluations of the generated products through comparison with other state-of-the-art methods confirm the validity of the approach, and open the door to operational applications at enhanced spatio-temporal resolutions at broad continental scales.

An agent-based model that simulates the spatio-temporal dynamics of sources and transfer mechanisms contributing faecal indicator organisms to streams. Part 2: Application to a small agricultural catchment.

The new Model for the Agent-based simulation of Faecal Indicator Organisms (MAFIO) is applied to a small (0.42 km2) Scottish agricultural catchment to simulate the dynamics of E. coli arising from sheep and cattle farming, in order to provide a proof-of-concept. The hydrological environment for MAFIO was simulated by the "best" ensemble run of the tracer-aided ecohydrological model EcH2O-iso, obtained through multi-criteria calibration to stream discharge (MAE: 1.37 L s-1) and spatially-distributed stable isotope data (MAE: 1.14-3.02‰) for the period April-December 2017. MAFIO was then applied for the period June-August for which twice-weekly E. coli loads were quantified at up to three sites along the stream. Performance in simulating these data suggested the model has skill in capturing the transfer of faecal indicator organisms (FIOs) from livestock to streams via the processes of direct deposition, transport in overland flow and seepage from areas of degraded soil. Furthermore, its agent-based structure allowed source areas, transfer mechanisms and host animals contributing FIOs to the stream to be quantified. Such information is likely to have substantial value in the context of designing and spatially-targeting mitigation measures against impaired microbial water quality. This study also revealed, however, that avenues exist for improving process conceptualisation in MAFIO (e.g. to include FIO contributions from wildlife) and highlighted the need to quantitatively assess how uncertainty in the spatial extent of surface flow paths in the simulated hydrological environment may affect FIO simulations. Despite the consequent status of MAFIO as a research-level model, its encouraging performance in this proof-of-concept study suggests the model has significant potential for eventual incorporation into decision support frameworks.

An agent-based model that simulates the spatio-temporal dynamics of sources and transfer mechanisms contributing faecal indicator organisms to streams. Part 1: Background and model description.

A new Model for the Agent-based simulation of Faecal Indicator Organisms (MAFIO) is developed that attempts to overcome limitations in existing faecal indicator organism (FIO) models arising from coarse spatial discretisations and poorly-constrained hydrological processes. MAFIO is a spatially-distributed, process-based model presently designed to simulate the fate and transport of agents representing FIOs shed by livestock at the sub-field scale in small (<10 km2) agricultural catchments. Specifically, FIO loading, die-off, detachment, surface routing, seepage and channel routing are modelled on a regular spatial grid. Central to MAFIO is that hydrological transfer mechanisms are simulated based on a hydrological environment generated by an external model for which it is possible to robustly determine the accuracy of simulated catchment hydrological functioning. The spatially-distributed, tracer-aided ecohydrological model EcH2O-iso is highlighted as a possible hydrological environment generator. The present paper provides a rationale for and description of MAFIO, whilst a companion paper applies the model in a small agricultural catchment in Scotland to provide a proof-of-concept.

Coupled ecohydrology and plant hydraulics modeling predicts ponderosa pine seedling mortality and lower treeline in the US Northern Rocky Mountains.

We modeled hydraulic stress in ponderosa pine seedlings at multiple scales to examine its influence on mortality and forest extent at the lower treeline in the northern Rockies. We combined a mechanistic ecohydrologic model with a vegetation dynamic stress index incorporating intensity, duration and frequency of hydraulic stress events, to examine mortality from loss of hydraulic conductivity. We calibrated our model using a glasshouse dry-down experiment and tested it using in situ monitoring data on seedling mortality from reforestation efforts. We then simulated hydraulic stress and mortality in seedlings within the Bitterroot River watershed of Montana. We show that cumulative hydraulic stress, its legacy and its consequences for mortality are predictable and can be modeled at local to landscape scales. We demonstrate that topographic controls on the distribution and availability of water and energy drive spatial patterns of hydraulic stress. Low-elevation, south-facing, nonconvergent locations with limited upslope water subsidies experienced the highest rates of modeled mortality. Simulated mortality in seedlings from 2001 to 2015 correlated with the current distribution of forest cover near the lower treeline, suggesting that hydraulic stress limits recruitment and ultimately constrains the low-elevation extent of conifer forests within the region.

DroughtCast: A Machine Learning Forecast of the United States Drought Monitor.

Drought is one of the most ecologically and economically devastating natural phenomena affecting the United States, causing the U.S. economy billions of dollars in damage, and driving widespread degradation of ecosystem health. Many drought indices are implemented to monitor the current extent and status of drought so stakeholders such as farmers and local governments can appropriately respond. Methods to forecast drought conditions weeks to months in advance are less common but would provide a more effective early warning system to enhance drought response, mitigation, and adaptation planning. To resolve this issue, we introduce DroughtCast, a machine learning framework for forecasting the United States Drought Monitor (USDM). DroughtCast operates on the knowledge that recent anomalies in hydrology and meteorology drive future changes in drought conditions. We use simulated meteorology and satellite observed soil moisture as inputs into a recurrent neural network to accurately forecast the USDM between 1 and 12 weeks into the future. Our analysis shows that precipitation, soil moisture, and temperature are the most important input variables when forecasting future drought conditions. Additionally, a case study of the 2017 Northern Plains Flash Drought shows that DroughtCast was able to forecast a very extreme drought event up to 12 weeks before its onset. Given the favorable forecasting skill of the model, DroughtCast may provide a promising tool for land managers and local governments in preparing for and mitigating the effects of drought.

Coastal development and precipitation drive pathogen flow from land to sea: evidence from a Toxoplasma gondii and felid host system.

Rapidly developing coastal regions face consequences of land use and climate change including flooding and increased sediment, nutrient, and chemical runoff, but these forces may also enhance pathogen runoff, which threatens human, animal, and ecosystem health. Using the zoonotic parasite Toxoplasma gondii in California, USA as a model for coastal pathogen pollution, we examine the spatial distribution of parasite runoff and the impacts of precipitation and development on projected pathogen delivery to the ocean. Oocysts, the extremely hardy free-living environmental stage of T. gondii shed in faeces of domestic and wild felids, are carried to the ocean by freshwater runoff. Linking spatial pathogen loading and transport models, we show that watersheds with the highest levels of oocyst runoff align closely with regions of increased sentinel marine mammal T. gondii infection. These watersheds are characterized by higher levels of coastal development and larger domestic cat populations. Increases in coastal development and precipitation independently raised oocyst delivery to the ocean (average increases of 44% and 79%, respectively), but dramatically increased parasite runoff when combined (175% average increase). Anthropogenic changes in landscapes and climate can accelerate runoff of diverse pathogens from terrestrial to aquatic environments, influencing transmission to people, domestic animals, and wildlife.