Varanasi Stock：Context-DEPENDENT AGRICULTURAL Intensification Pathways to Increase Rice Production in India

The Research Conductational Herein Was Review by and Complies with Standards Established by the Research Ethics Committee of the International Maize and WHEAT VEMENT Center (Cimmyt) as described in Policy Number DDG-POL-04–2019..2019.06.Varanasi Stock

The Study Area Comprises The Seven Major Rice Production States of India, namely Eastern UTTAR PRADESH and BIHAR (n = 10,714 Field-year Combinings) 47), jharkhand (n = 717), chhattisgarh (n = 1099),West Bengal (n = 1363), and andHra Pradesh (n = 1046) during the 2017, 2018, and 2019 Monsoon Seasons (Supplementary Fig. 5). PS. FIRST, AttaIINable Yield Gaps (YGA) WEREESTIMATED for Each State. Second, Random Forest Analytics Were Developed to Identify the Most Important Variables Explaining Yield StateJaipur Investment. Thing SHA Pely addictive Explanation (Shap) WAS Used to Segregate the Relate Contribution of Each Prational to Rice Yield Prediction and, Finally, Machine Learning-Based SCENARIO Analysis was used to quantify the beneFits of Integrated CropMANAGEMENT PRACTIES At Regional Level, With A Focus I n bihar and adjacent areas of Eastern Uttar Pradesh where data collection.in A Spatial Context for Sustainable Rice Intensification in the Region Through A Geographical Hotspot Analysis.

The landscape-scale crop assessment surveys were conducted with digital collection tools and requested information on agronomic management practices and biophysical characteristics for the largest rice production field in each farm. The farmer reported yield was verified by measured crop cut yield through harvesting a 2 × 2M Quadrant Randomly from the Repressentative Center of the FROM A Fraction of Farms. Survey Data and the Corresponding Data Collection Tool Arely AvaiLely onLely He Details of the Sampling and Data Collection Protocols Are Reported Elsewhere12. Survey Data for Each Field was then CombinedWith Gridded Daily Weather Data from the Reported Sowst Dates from Nasa Power (). Descriptive Statistics of the Variables used in the Analysis are d in supportmentary table 2.

The Attainable Yield Gap (YGA) WAS ESTIMATED As The Differentene The Mean Actual Yield Yield Yield Yielding Fields (I.E., Top 10 Percentile of the Yield DI Stribution; The Attainable Yield) in Each State and the Actual Yield Observed in All Other Fields in theRespective state. TheReaFTER, State-Specific Random Forest Models Weeeels, and FINE-TUNED FOLLOWING NAYAK et al.14, To Identify the Most Important Factors E XPlaining Rice Yield Variability in Each State. Model Over-Fitting was avoided by Keeping at Least 50observations in all terminal nodes with each time.

After Each Model Was Built, Individual Conditional Expectance (Ice) Plots48 WERE CREATED for The Two Most Important Practices, Sequentially, As Identi FIED BY PERMUTATATION-BASED Feature Importance. Ice Plots Were Developed for Individual Fields to Predict The Relationship BetWeen The Most Important Input InputVariables and Rice Yield, While Keeping All Other Input Variables at their Reported Value for Each Field. Suppose N Fertilizer Rate Was Identified As TH E Most Important Varialization to Explain Rice Yield Variability, then Crop Yield was Predict with Ice for a Vector of aN application rating the range of n application raters observation in the data in steps of 10 kg n ha−1. N application rate and the maximum yield across the vector of n application rate (YSTEP1)Reflects The Expected Yield Gap CLOSURE ONCE The MOST Important Constraint is Addressed (YG1). Subsequently, The Original Feature Values ​​(I.E., N Application e, in this example) for Each Field Wee REPLACED with the Corresponding N Application Rate Association with the Maximum YieldFrom the Ice (nyld_max) and Ice Plots Wee Created Again for the Second Most Important Variable in Combination with Nyld_max and The Reported Values ​​Used in the model for each fire (ytep2). The Differentce Between The YSTEP2 and YSTEP1 is definedAs the Expected Yield Gap Closure once The Second Most Important Constraint is Addressed (YG2). Ield Gap Closure after Removing Yield Constraints Association with the Two Most Important Varios Explaining Yield Variability.

The distribution of yg1 and yg2, and their sum, was expressed related, by the two most important management prACTices. The yield gap analysis was confined with the ranger and caret R Packages49, 50 And with the iml r package51.

A Shapley Additive Explanation (Shap) -Based Methodology was further deployed to quantify the relatives of biophysical faction. Actices to Rice Yields Prediction for 10,714 Field-Year Combinings in Bihar and Adjacent DistricTs in Eastern UTTAR PRADESH.Shap is a post-HOC Methodology to Interpret Random Forest Models and Identify Heterogeneous Effect of Management Practices on Yield Prediction. Operating Game theory and Is used to enter the marginal contribution of each player to a team's overall performance52. by Conceptualization IndividualFields as a 'Team' And Management Practices as 'Players', The Shap Methodology Can be used to quantify the relatives Ice to crop yield outComes on Individual Field. Hence, with this method method, the contextual value of disferingCan be transitived into Pathways for increasing crop props.

Shap is an additive feature attribution method that is used as a post-Hoc approach for local interpretation of data-driven models. In Short Tion of Individual Variables to Model Predictions of the Outcom of Interest52. Thus, The Shap Value forAny Variable J (ϕj; T Ha−1) Can be interpreted in our Assessment as the marginal control. TED RICE YIELD Across the DataSet. In Other Words, Shap Values ​​Refer to TheYield Contribution of Individual Variables, Expressed As Either a Positive or Negative Deviation from the Population Mean. VE Shap Values ​​in Absolute Terms Have a Large Positive and Negative Influence on Modeled Yield Predictions, Respectively.Each Field with the IML R PACKAGE51 and Visualized In Relation to the Scaled Absolute Valiable Values. Numeric Variable Normalized USING Scaling, and Categorical Variables WeRDinally Factored and Scaled (Supplementary Table 1). The Spatial distribution of the two managementWith Highest Absolute Shap Value in Bihar and Eastern Uttar Pradessed Through A Hot Spot SPOT SPOTSIS CONDUCTD In Arcgis Pro 2.9.0 And Consisting of T He Calculation of the Getis-Ord GI* Statistics for Each Field Assump a Fixed Distance Band of 10 KM.

Absolute shap value for each input variable weereaged across fields to rank most toworthtantant level. Four Clusters Based on the Two Most Important Management Practices at State Level to Delineate Where A Single Versus Multiple Production PracticeCHANGES Are Import to Increase Rice Productivity. The Clustering Analysis SUPPPPORTS A TARGETD Approach to Sustainable Intensification, As Opposes PProach WHERE All Fields Receive The Same Intervention. The Clustering Was Done Based On the Shap Value of IRRIGATIONS and N Fertilizer RateFor Each Field. The Clusters Included Fields with A Positive Shap Value for Both Practices (I+N+), a Negative Shap Value for Both Practices (I−N−), and A PO SITIVE Shap Value for One and A Negative Shap Value for TheOther Practice (I+N− and I−n+). As Such, Positive Shap Values ​​for a Given Management Practure Can be considerRed, s more yield limiting, indicating when improved management can generate yield gains.

Four Sustainable Intensification SCENARIOS WeRe Designed to Explore the Aggregated Production Benefits, Additional Input Requirements, and Profitability of Yi Eld Gap Closure in Bihar and Eastern Uttar Pradesh as Compared to Current Farmers' Practice. SCENARIO 1 CONSISTED of the State Level Blanket of 125 kg nhal1 in All Fields. SCENARIO 2 Consist of Blanket Use of 180 KG N HA−1, Defined Based on the PARTIAL DEPENDENCY PLA RANDOM FOREST MODEL FITTED To the POOLED DATA FA OR bihar and Eastern UTTAR PRADESHMumbai Wealth Management. TheRe Wee Two Cluster-Based Targeting SCENARIOS. SCENARIO 3 Considered Interventions for The N Fertilizer Only in the I+N− and I−n− CluusttersAhmedabad Stock. Ese Clusters, whereas n raters we not changed inThe Other Clusters. SCENARIO 4 Consisted of Cluster-Based Targeting for Both N and Irrigation, I.E., Addressing The Co-Limitation Factors In A Limite. d number of fars. The n use and Irigation in I−n− Cluster was changed to180 kg n hak1 and 5 IRRIGATIONS. Crop Management Remaind Unchanged in the Other Clusters. For SCENARIOS 1 and 2, The aggregated benefit in terms of additional Tion was Obtained by Multiplying The Predictid Yield Increases (T HA−1) by Total RiceArea of ​​Each distribution. For SCENARIOS 3 and 4, The Additional Rice Yield, Additional Water and N Use, and RETURNS on inverted (I.E., Additive Resources, $ 20 USD Per IRrication and $ 0.14 USD PER KG Subsidized N, Subtracted from AdDitional Rice SalesRevenues Based on the Rice Minimum SUPPPORT Price for 2018) WERE ESTIMATED At the District Leveling The Share of Farms in Each Cluster where interventions d.

All Data Analysis Were Conducting in R (4.2.3) with The Follow Package and Version Number, DPLYR (1.1.4), Caret (6.0.93), Range (0.14.1), IML (0.11.1), Ge Odata (Odata (0.5.3), Terra (1.7.55), Tidyverse (1.3.2), ggpubr (0.6.0) and dependencies, data.table (1.14.2), and gridextra (2.3). t in arcgisPro v.2.9.0.Hyderabad Stocks

FURTHER Information on Research Design is available in the nature portfolio reporting summary Linked to this article.

Mumbai Wealth Management