Extrapolation Domains Research Articles

Pathways and determinants of changing nutrient management in lowland rice-based systems of Southeast Asia

Production increases in lowland rice-based systems of Southeast Asia will not be possible without the addition of nutrients. Recent productivity gains were largely based on increasing the use efficiency of the applied nutrients through integrated and site-specific approaches. The adoption of such strategies has evolved differentially in the various country. We need to understand such recent trends, their driving forces, and their effectiveness to harness or leverage them for regional food security and sustainable development. To gain this understanding, we analyzed changes in fertility-related agronomic practices in six representative rice-based production systems of Southeast Asia (three favorable and three marginal sites). We implemented a diachronic survey (comparing the years 2000 and 2018) combined with field sampling and measurements in 1024 rice-producing households and, where applicable, in both dry and wet seasons (3638 complete data sets). We observed that soil fertility management followed similar trends across sites. However, the timing and the extent of changes were location specific, and differed primarily by the favorability/marginality of the site, and farmers’ capability to adopt technological innovations. The irrigated rice double cropping benefitted most from recent technology innovations (hybrid seeds, farm machinery), accelerating the adoption of improved fertility management practices to an aggregated level of up to 75% (mean across the six soil fertility-related management practices) at favorable sites. Most prominent were multiple splitting of mineral N and the application of mineral fertilizers other than N (P, K, Zn). At marginal sites, aggregate changes toward intensification were rather low with 35–55%, and farmers tended to intensify the non-rice component in the system by applying mineral fertilizers and organic amendments to dry season upland crops. A balanced use of nutrients will require the targeting of options to specific extrapolation domains (social-ecological target environments) based on site attributes and farmers’ resource endowment.

Evaluating the Impact of Scaling Considering the Extrapolation Domain on the Prediction Performance of Machine Learning Algorithms

In this study, we used a benchmark dataset to evaluate the impact of scaling with the extrapolation domain on the prediction performance of machine learning algorithms. We pseudo-divided the data into the interpolation domain (training data) and the extrapolation domain (test data) using a combination of UMAP (Uniform Manifold Approximation and Projection) and material domain knowledge. In anticipation of bridging interpolation and extrapolation domains in nonlinear machine learning algorithms, we evaluated how the scaling considering the extrapolation domain affects prediction performance in the extrapolation domain. For this evaluation, we used three nonlinear algorithms widely used in the MI (Materials Informatics) domain: XGB (XGBoost) regression, GP (Gaussian Process) regression, and SVR (Support Vector Regression). In this study, by defining the pseudo extrapolation domain, we established the approach for evaluating the prediction accuracy of machine learning models in the extrapolation domain, which is considered difficult to evaluate quantitatively. We also demonstrated that this method, which uses scaling that considers the extrapolation domain, is an effective method for improving prediction accuracy in the extrapolation domain while maintaining prediction accuracy in the interpolation domain.

Crop-specific technology extrapolation domains for Brazil

ABSTRACT Climate and soil databases can be used for delineating homogenous environmental zones over large areas of agricultural land to support agricultural research planning and spatial extrapolation of the results of field experiments. In other countries, such framework has proven useful to improve the efficiency and impact of investments in agricultural research. Furthermore, this approach can also contribute to identify key locations with the greatest potential for spatial extrapolation in terms of crop production area with similar climate and soils. In this paper, soil and weather databases covering the main agricultural land in Brazil were developed, and crop-specific technology extrapolation domains (TEDs) were identified for maize, soybean, and sugarcane crops. These databases were useful and viable option for the generation of homogeneous zones maps for different crops and regions. Given the spatial scale of the database applied, it was possible to represent a large area in few TEDs, without losing information regarding climate and soil variability, facilitating the spatial extrapolation of technologies throughout Brazil. Few TEDs are needed to represent a large part of the area of sugarcane (32 TEDs to cover 51% of the area), maize (21 TEDs for 51% of the area), and soybean (17 TEDs for 50% of the total area). Such information may help several areas of the agricultural sector, such as supporting new research that works on areas with large scales, or areas not explored yet.

Nitrogen response functions targeted to technology extrapolation domains in Ethiopia using CERES‐maize

AbstractThe profitability of fertilizer‐N use can be optimized using N response functions specific to climate‐based technology extrapolation domains (TED). Crop growth simulation can complement field research for targeting of response functions. The objective of this study was to target maize (Zea mays L.) N response functions to seven TED in Ethiopia through CERES‐Maize simulation of continuous maize over 30 yr. The complete factorial set of treatments included seven levels of N in 25 kg ha−1 increments under no‐till (NT) and conventional tillage (CT) systems. The CERES‐Maize simulated experiments were done for two or three sites per TED. Nitrogen response functions were generated for each TED with tillage‐specific functions for three TED with tillage × N interactions. The N rate responses for all TED fit curvilinear to plateau functions but with differing magnitudes and shapes of response. The mean yield with NT was 6% less than with CT, but the difference declined with increased N rate. The economically optimum N rate (EONR) ranged from 65 to 179 and 103 to 243 kg ha−1 for high and low‐cost fertilizer‐N, respectively. The EONR was 6% less and the profit cost ratio was 11% higher with CT compared to NT, indicating greater fertilizer‐N need with NT than with CT. The application of N for maize was highly profitable for all TED. The EONR from CERES‐Maize were higher than past field research results. This suggests that the CERES‐Maize N response functions were most appropriate for well‐managed crop production situations in Ethiopia.

Combining field-level data and remote sensing to understand impact of management practices on producer yields

Abstract Producer field-level data have been used in recent studies to identify suites of management practices that consistently increase yield for a given climate-soil environment. However, the physiological drivers underlying the empirical associations between yield and management practices derived from these studies have remained mostly speculative, particularly in terms of resource capture and conversion into economic yield. We followed an approach consisting of a large producer database, satellite imagery, and crop modeling to assess which key physiological parameters best explain on-farm yield response to producer-chosen management practices across different climate-soil production environments. Survey data on yield and management practices were collected from 5291 soybean fields sown across the US North Central region during 2014-2016. Producer data were grouped into 10 technology extrapolation domains (TEDs) to account for variation in soil, climate, and water regime across producer fields. Simulated phenology and satellite imagery were used to estimate incident (IPAR) and absorbed (APAR) photosynthetically active radiation during the entire crop season and also during the time-span beginning with pod setting and ending with physiological maturity. In each TED, seed yield was increased by early sowing (+13 to +39 kg ha-1d-1), which was physiologically driven by increases in the duration of both crop cycle and critical period, and by a higher APAR. In-season application of foliar fungicide and/or insecticide also resulted in significant yield increase (+0.35 Mg ha−1, p

Assessing benefits of artificial drainage on soybean yield in the North Central US region

Artificial drainage (AD) in producer fields can help avoid excess water and improve workability and timely fieldwork in comparison with soils that rely on natural drainage (ND). To date, most studies examining the effect of AD on crop yield were conducted at research stations and limited to a few sites and years. Here, we explored the influence of AD on soybean yield across the North Central US region (NC-US), which accounts for a third of global soybean production and where AD is widely used to prevent excess water early in the crop season. We used two sources of data: (i) records from 2805 soybean producer fields collected via a multi-year, multi-state survey of soybean producers in the NC-US region; and (ii) information from 47 site-year experiments that included paired AD-ND treatment comparisons. In all cases, AD corresponded to subsurface drainage, except for producer fields in North Dakota where AD corresponded to surface drainage. Producer fields were grouped into technology extrapolation domains (TEDs) delineating regions with similar climate and soils to allow comparison of yields in AD versus ND fields. In the case of subsurface drainage, average yield in AD versus ND was 8% higher (+275 kg ha−1) and 4% higher (+157 kg ha−1) based on analysis of experimental and producer data, respectively. Our analysis indicated that part of the AD-ND yield difference in producer fields can be attributed to a shift towards earlier sowing after AD adoption. Findings from this study explain the wide adoption of AD across the central and eastern areas of the NC-US region and provide a basis to determine the productivity and economic impact of AD installation at field and regional level.

Sustainable Agroforestry Landscape Management: Changing the Game

Location-specific forms of agroforestry management can reduce problems in the forest–water–people nexus, by balancing upstream and downstream interests, but social and ecological finetuning is needed. New ways of achieving shared understanding of the underlying ecological and social-ecological relations is needed to adapt and contextualize generic solutions. Addressing these challenges between thirteen cases of tropical agroforestry scenario development across three continents requires exploration of generic aspects of issues, knowledge and participative approaches. Participative projects with local stakeholders increasingly use ‘serious gaming’. Although helpful, serious games so far (1) appear to be ad hoc, case dependent, with poorly defined extrapolation domains, (2) require heavy research investment, (3) have untested cultural limitations and (4) lack clarity on where and how they can be used in policy making. We classify the main forest–water–people nexus issues and the types of land-use solutions that shape local discourses and that are to be brought to life in the games. Four ‘prototype’ games will be further used to test hypotheses about the four problems identified constraining game use. The resulting generic forest–water–people games will be the outcome of the project “Scenario evaluation for sustainable agroforestry management through forest-water-people games” (SESAM), for which this article provides a preview.

Management strategies for early‐ and late‐planted soybean in the north‐central United States

AbstractIt is widely recognized that planting soybean [Glycine max (L.) Merr.] early is critical to maximizing yield, but the influence of changing management factors when soybean planting is delayed is not well understood. The objectives of this research were to (a) identify management decisions that increase seed yield in either early‐ or late‐planted soybean scenarios, and (b) estimate the maximum break‐even price of each management factor identified to influence soybean seed yield in early‐ or late‐planted soybean. Producer data on seed yield and management decisions were collected from 5682 fields planted with soybean during 2014−2016 and grouped into 10 technology extrapolation domains (TEDs) based on growing environment. A subsample of 1512 fields was classified into early‐ and late‐planted categories using terciles. Conditional inference trees were created for each TED to evaluate the effect of management decisions within the two planting date timeframes on seed yield. Management strategies that maximized yield and associated maximum break‐even prices varied across TEDs and planting date. For early‐planted fields, higher yields were associated with artificial drainage, insecticide seed treatment, and lower seeding rates. For late‐planted fields, herbicide application timing and tillage intensity were related to higher yields. There was no individual management decision that consistently increased seed yield across all TEDs.

Improved soil biological health increases corn grain yield in N fertilized systems across the Corn Belt

Nitrogenous fertilizers have nearly doubled global grain yields, but have also increased losses of reactive N to the environment. Current public investments to improve soil health seek to balance productivity and environmental considerations. However, data integrating soil biological health and crop N response to date is insufficient to reliably drive conservation policy and inform management. Here we used multilevel structural equation modeling and N fertilizer rate trials to show that biologically healthier soils produce greater corn yields per unit of fertilizer. We found the effect of soil biological health on corn yield was 18% the magnitude of N fertilization, Moreover, we found this effect was consistent for edaphic and climatic conditions representative of 52% of the rainfed acreage in the Corn Belt (as determined using technological extrapolation domains). While N fertilization also plays a role in building or maintaining soil biological health, soil biological health metrics offer limited a priori information on a site’s responsiveness to N fertilizer applications. Thus, increases in soil biological health can increase corn yields for a given unit of N fertilizer, but cannot completely replace mineral N fertilization in these systems. Our results illustrate the potential for gains in productivity through investment in soil biological health, independent of increases in mineral N fertilizer use.

Benchmarking impact of nitrogen inputs on grain yield and environmental performance of producer fields in the western US Corn Belt

Benchmarking crop yields against nitrogen (N) input levels can help provide opportunities to improve N fertilizer efficiency and reduce N losses on maize in the US Corn Belt by identifying fields most likely to benefit from improved N management practices. Here, we evaluated a large producer database that includes field-level data on yield and applied N inputs from 9280 irrigated and rainfed fields over a 7-year period (2009–2015) in Nebraska (USA). A spatial framework, based on technology extrapolation domains, was used to cluster each field into spatial units with similar climate and soil type that represent 1.3 million ha of US farm land sown annually with maize. Three metrics were employed to evaluate agronomic and environmental performance: partial factor productivity for N inputs (PFPN, ratio between yield and N inputs), N balance (difference between N inputs and grain N removal), and yield-scaled N balance (ratio between N balance and yield). Nitrogen inputs included N from fertilizer and N contained in applied irrigation water. Average yield and N inputs were 40 and 44% higher in irrigated versus rainfed fields. The N balance was ca. 2-fold greater in irrigated versus rainfed fields (81 versus 41 kg N ha−1). Of the total number of field-years, 58% (irrigated) and 15% (rainfed) had N balance ≥ 75 kg N ha−1, which was considered a threshold to identify fields with potentially large N losses. Very large (> 150 kg N ha−1) and negative N balance estimates were not apparent when analysis was based on field averages using a minimum of three years' data instead of individual field-years. Nitrogen balance was smaller for maize crops following soybean compared to continuous maize. Despite the larger N balance (on an area basis), irrigated fields exhibited smaller yield-scaled N balance relative to rainfed fields. The approach proposed here can readily be adopted to benchmark current use of N fertilizer for other cereal-based crop systems, inform policy, and identify opportunities for improvement in N management.

GIS-Based multi-criteria land suitability mapping for scaling faba bean varieties in Ethiopia

Successful scaling of agricultural technology requires a spatial explicit framework for targeting the right variety at the right place. This entails a multi-criteria evaluation (MCE) approach, using a set of determining factors to delineate the scaling domains for faba bean (Vicia faba L.) varieties in and identify potentially suitable land area in a specific region, and zone in Ethiopia. Meeting this challenge will require a solid spatial framework. Land suitability analysis is an evaluation and spatial decision making, involving several determining factors. The factors considered in this analysis include key biophysical parameters such as climate, topography, soil types and properties. The analysis was also focused on improved faba bean varieties viz., Dagm, Dosha, Gabelcho, Gora, Hachalu, Moti and Walki. The environmental factors’ layers of a specific crop pixel values were classified and given a weight, and then compared among themselves for further ranking to account for their relative importance to delineate variety specific extrapolation domains. The geo-statistical analysis was carried out to estimate the extent of the scalable areas. The classification showed that, it was highly suitable for varieties 0.02 million hectares for Dosha; 0.19 for Gabelcho; 0.11 for Gora; 0.33 for Moti; 0.05 for Dagm; 0.14 for Hachalu; and 0.26 million hectares for Walki. Moderately suitable areas for these varieties covered 5.0, 9.4, 7.2, 15.3, 4.6, 8.8, and 7.5 million hectares, respectively across the country. The largest proportion for all varieties was moderately suitable; while the share of slightly suitable was very low, although there was quite variability within each of the faba bean variety in terms of its agro-ecology adaptation to the target environments. Such biophysical spatial frameworks become essential entry points for introducing variety specific product profiles and this can be further enhanced by incorporating socio-economic attributes accounting for return of the investment in targeting the technology.Key words: Environmental factors, spatial decision

Geospatial modeling of conservation tillage and nitrogen timing effects on yield and soil properties

Crop growth simulation can complement field research for adapting and targeting practices to diverse production areas. Two simulated experiments of 30-year duration were conducted with CERES and CROPGRO to evaluate the effects of fertilizer N practices and conservation tillage (CT) alternatives on maize (Zea mays L.) grain yield, soil organic C and soil organic N for seven technology extrapolation domains (TED) in Ethiopia. Mean maize grain yield was 663kg ha−1 more with three compared two N applications per season for high rainfall TED in western Ethiopia. Tillage did not affect response to N. Averaged across TED, maize yield was 33% more with a combination of conservation tillage, rotation and additional N application (CTr+N) compared maize monoculture with conventional tillage and the recommended N rate (CPmm), primarily because of crop rotation. Maize grain yield increased over time with the conservation tillage under rotation (CTr) but declined under CPmm. Soil organic C and N declined over time, but the rate of decline was lower with CTr+N compared to CPmm. Stored soil organic C and N were 8543 and 594kgha−1 more with CTr+N compared with CPmm, respectively, averaged over the 30 years. First order stochastic dominance analysis from maize grain partial net returns showed that maize rotation dominated maize monoculture. Second order stochastic dominance analysis for partial net return indicated the absence of an unambiguous dominant of conservation tillage for a farmers’ preference of low risk to higher net return or high net return to low risk.

A spatial framework for ex-ante impact assessment of agricultural technologies

Traditional agricultural research and extension relies on replicated field experiments, on-farm trials, and demonstration plots to evaluate and adapt agronomic technologies that aim to increase productivity, reduce risk, and protect the environment for a given biophysical and socio-economic context. To date, these efforts lack a generic and robust spatial framework for ex-ante assessment that: (i) provides strategic insight to guide decisions about the number and location of testing sites, (ii) define the target domain for scaling-out a given technology or technology package, and (iii) estimate potential impact from widespread adoption of the technology(ies) being evaluated. In this study, we developed a data-rich spatial framework to guide agricultural research and development (AR&D) prioritization and to perform ex-ante impact assessment. The framework uses “technology extrapolation domains”, which delineate regions with similar climate and soil type combined with other biophysical and socio-economic factors that influence technology adoption. We provide proof of concept for the framework using a maize agronomy project in three sub-Saharan Africa countries (Ethiopia, Nigeria, and Tanzania) as a case study. We used maize area and rural population coverage as indicators to estimate potential project impact in each country. The project conducted 496 nutrient omission trials located at both on-farm and research station sites across these three countries. Reallocation of test sites towards domains with a larger proportion of national maize area could increase coverage of maize area by 79–134% and of rural population by 14–33% in Nigeria and Ethiopia. This study represents a first step in developing a generic, transparent, and scientifically robust framework to estimate ex-ante impact of AR&D programs that aim to increase food production and reduce poverty and hunger.

Beyond the plot: technology extrapolation domains for scaling out agronomic science

Ensuring an adequate food supply in systems that protect environmental quality and conserve natural resources requires productive and resource-efficient cropping systems on existing farmland. Meeting this challenge will be difficult without a robust spatial framework that facilitates rapid evaluation and scaling-out of currently available and emerging technologies. Here we develop a global spatial framework to delineate ‘technology extrapolation domains’ based on key climate and soil factors that govern crop yields and yield stability in rainfed crop production. The proposed framework adequately represents the spatial pattern of crop yields and stability when evaluated over the data-rich US Corn Belt. It also facilitates evaluation of cropping system performance across continents, which can improve efficiency of agricultural research that seeks to intensify production on existing farmland. Populating this biophysical spatial framework with appropriate socio-economic attributes provides the potential to amplify the return on investments in agricultural research and development by improving the effectiveness of research prioritization and impact assessment.

Sifting and winnowing: Analysis of farmer field data for soybean in the US North-Central region

Field trials are commonly used to estimate the effects of different factors on crop yields. In the present study, we followed an alternative approach to identify factors that explain field-to-field yield variation, which consisted of farmer survey data, a spatial framework, and multiple statistical procedures. This approach was used to identify management factors with strongest association with on-farm soybean yield variation in the US North Central (NC) region. Field survey data, including yield and management information, were collected over two crop growing seasons (2014 and 2015) from rainfed and irrigated soybean fields (total of 3568 field-year observations). Fields were grouped into technology extrapolation domains (TEDs) that accounted for soil and climate variation and 9 TEDs were selected based on the number of fields needed to detect yield differences due to management as determined using power analysis. Average yield ranged from 2.5 to 5 Mg ha−1 across TEDs, with field yield distributions in half of the domains having a distributional peak that was close to maximum yields. Conditional inference trees analysis was chosen among 26 statistical procedures as the approach that best combines ability to detect and rank factors (and their interactions) with greatest influence on on-farm yield and relatively easy interpretation of results. Survey data from ca. 150 fields in each of the nine TEDs allowed us to identify key management factors influencing yields for an agricultural area that includes ca. 7 million ha sown with soybean. In five of the nine TEDs, highest yields were observed in early-sown fields. Other factors explaining on-farm yield variation were maturity group, and in-season foliar fungicide and/or insecticide application, but, in some cases, their influence on yield depended upon sowing date and water regime. While the approach proposed here cannot establish cause-effect relationships conclusively, it can certainly provide a focus to replicated field experiments in relation to which management factors to investigate. We believe that future agronomic studies based on farmer survey data can greatly benefit from ex-ante identification of most important TEDs (relative to crop area and production) as well as determination of minimum number of farmer survey data that needs to be collected from each of them based on expected yield differences and variability. The approach is generic enough to be applied in other crop producing regions as long as farmer data and associated climate and soil databases are available.

GEOSPATIAL APPROACH FOR DELINEATING EXTRAPOLATION DOMAINS FOR SUSTAINABLE AGRICULTURAL INTENSIFICATION TECHNOLOGIES

Abstract. Sustainable intensification (SI) is a viable pathway to increase agricultural production and improve ecosystem health. Scaling SI technologies in locations with similar biophysical conditions enhance adoption. This paper employs novel extrapolation detection (ExeDet) algorithm and gridded bioclimatic layers to delineate extrapolation domains for improved maize variety (SC719) and inorganic fertilizers (YaraMila-CEREAL® and YaraBela-Sulfan®) in Tanzania. Suitability was based on grain yields recorded in on-farm trials. The ExeDet algorithm generated three maps: (1) the dissimilarity between bioclimatic conditions in the reference trial sites and the target extrapolation domain (Novelty type-1), (2) the magnitude of novel correlations between covariates in extrapolation domain (Novelty type-2) and (3) the most limiting covariate. The novelty type1 and 2 maps were intersected and reclassified into five suitability classes. These classes were cross-tabulated to generate extrapolation suitability index (ESI) for the candidate technology package. An impact based spatial targeting index (IBSTI) was used to identify areas within the zones earmarked as suitable using ESI where the potential impacts for out scaling interventions can be maximized. Application of ESI and IBSTI is expected to guide extension and development agencies to prioritize scaling intervention based on both biophysical suitability and potential impact of particular technology package. Annual precipitation was most limiting factor in largest area of the extrapolation domain. Identification of the spatial distribution of the limiting factor is useful for recommending remedial measures to address the limiting factor that hinder a technology to achieve its full potential. The method outlined in this paper is replicable to other technologies that require extrapolation provided that representative reference trial data and appropriate biophysical grids are available.

Assessing causes of yield gaps in agricultural areas with diversity in climate and soils

Identification of causes of gaps between yield potential and producer yields has been restricted to small geographic areas. In the present study, we developed a novel approach for identifying causes of yield gaps over large agricultural areas with diversity in climate and soils. This approach was applied to quantify and explain yield gaps in rainfed and irrigated soybean in the North-Central USA (NC USA) region, which accounts for about one third of soybean global production. Survey data on yield and management were collected from 3568 producer fields over two crop seasons and grouped into 10 technology extrapolation domains (TEDs) according to their soil, climate, and water regime. Yield potential was estimated using a combination of crop modeling and boundary functions for water productivity and compared against highest producer yields derived from the yield distribution in each TED-year. Yield gaps were calculated as the difference between yield potential and average producer yield. Explanatory factors for yield gaps were investigated by identifying management practices that were concordantly associated with high- and low-yield fields. Management×TED interactions were then evaluated to elucidate the underlying causes of yield gaps. The chosen spatial TED framework accounted for about half of the regional variation in producer yield within the NC USA region. Across the 10 TEDs, soybean average yield potential ranged from 3.3 to 5.3 Mgha−1 for rainfed fields and from 5.3 to 5.6Mgha−1 for irrigated fields. Highest producer yields in each TED were similar (±12%) to the estimated yield potential. Yield gap, calculated as percentage of yield potential, was larger in rainfed (range: 15–28%) than in irrigated (range: 11–16%) soybean. Upscaled to the NC USA region, yield potential was 4.8Mgha−1 (rainfed) and 5.7Mgha−1 (irrigated), with a respective yield gap of 22 and 13% of yield potential. Sowing date, tillage, and in-season foliar fungicide and/or insecticide were identified as explanatory causes for yield variation in half or more of the 10 TEDs. However, the degree to which these three factors influenced producer yield varied across TEDs. Analysis of in-season weather helped interpret management×TED interactions. For example, yield increase due to advances in sowing date was greater in TEDs with less water limitation during the pod-setting phase. The present study highlights the strength of combining producer survey data with a spatial framework to measure yield gaps, identify management factors explaining these gaps, and understand the biophysical drivers influencing yield responses to crop management.

Tropical forest-transition landscapes: a portfolio for studying people, tree crops and agro-ecological change in context

ABSTRACTNudging the development trajectory of tropical landscapes towards sustainability requires a global commitment and policies that take diverse contexts and forest transitions into account. Out-scaling and upscaling landscape-level actions to achieve sustainable development goals globally need to be based on understanding of extrapolation domains and interconnectivity of products and services. We evaluated three portfolios of tropical landscape observatories and quantified extrapolation domains across ecological zones, stages of forest transition, human development index (HDI), population density and potential prominence of four dominant tropical tree crops (arabica coffee, cacao, rubber and oil palm). The ASB Partnership for Tropical Forest Margins portfolio was focussed on active humid forest margins and the Poverty and Environment Network on early stages of forest transition. The portfolio of sentinel landscapes of the Forests, Trees and Agroforestry (FTA) research programme provides a 5% sample of pantropical area, 8% of people, 9% of tree cover and 10–12% of potential tree crop presence, with quantified biases across zones, transition stages and HDI. In the ‘water tower’ configuration, relatively high population density coincides with biodiversity, coffee expansion and contested ecosystem services. The extrapolation domain of the FTA portfolio includes trade-off (tree loss) and synergy (restoration) phases of tropical forest transition.EDITED BY Paolo Omar Cerutti

Can crop-climate models be accurate and precise? A case study for wheat production in Denmark

Crop models, used to make projections of climate change impacts, differ greatly in structural detail. Complexity of model structure has generic effects on uncertainty and error propagation in climate change impact assessments. We applied Bayesian calibration to three distinctly different empirical and mechanistic wheat models to assess how differences in the extent of process understanding in models affects uncertainties in projected impact. Predictive power of the models was tested via both accuracy (bias) and precision (or tightness of grouping) of yield projections for extrapolated weather conditions. Yields predicted by the mechanistic model were generally more accurate than the empirical models for extrapolated conditions. This trend does not hold for all extrapolations; mechanistic and empirical models responded differently due to their sensitivities to distinct weather features. However, higher accuracy comes at the cost of precision of the mechanistic model to embrace all observations within given boundaries. The approaches showed complementarity in sensitivity to weather variables and in accuracy for different extrapolation domains. Their differences in model precision and accuracy make them suitable for generic model ensembles for near-term agricultural impact assessments of climate change.

Strategic parameter search method based on prediction errors and data density for efficient product design

In experimental design for functional molecules, materials and products, complex relationships exist between the various experimental parameters and the target physical and chemical properties. Regression analyses with experimental data are a useful way to understand these relationships. A carefully constructed regression model can be used to search for functional products effectively. However, although these products can be found in the extrapolation domains of the existing data, the predictive ability of the model tends to be low in regions where the data density is low, and new candidates where the predicted values of a property are unreliable will not achieve the desired property values. Therefore, to search for new candidates in appropriate extrapolation domains, we consider the probability that a new candidate will have the intended property values and the reliability of a predicted property value for this candidate. The probability is calculated from a predicted value and the estimated prediction error, and the reliability is based on the data density. The proposed method is applied to both simulation data and aqueous solubility data, and the efficiency of the method is demonstrated through data analyses.