My Notes on Sustainable Development

15 DATA-RICH BRIEFS ON SDG POLICY, COMMUNITIES AND APPLIED SCIENCE

GLOBAL DELIVERY 3 stories

JOBS & PROTECTION 5 stories

COMMUNITY & STEM 5 stories

ISSUE DATE19 July 2026

CONTENTS

GLOBAL GOVERNANCE AND SDG DELIVERY

01  HLPF 2026 Moves the SDG Debate from Separate Targets to Integrated Delivery

02  Local2030 Puts Cities at the Centre of SDG Delivery—but Finance and Data Remain the Test

03  The Cost of a Healthy Diet Is Now a Macroeconomic and Social-Protection Problem

JOBS, SKILLS AND SOCIAL PROTECTION

04  Paraguay’s Job Growth Is Substantial—Job Quality and Attribution Are the Next Questions

05  Albania’s LEAP Programme Connects Training to Jobs

06  Uzbekistan’s Social-Insurance Law to  Reach 5.8 Million More People

07  Malawi’s Compensation Reform Could Protect 4.2 Million Workers

08  Cambodia’s Skills Programme Has Reached 271,000 Enrolments

COMMUNITY-LED ADAPTATION AND CONSERVATION

09  Atacama Communities Turn Fog into a Supplementary Water Source

10  Palawan’s Turtle Patrollers Convert Local Knowledge into Conservation Capacity

11  Vodun Sacred Zones Give Benin’s Mangroves a Locally Enforced Protection System

12 ‘Bear-dar’ Demonstrates One Successful Early Warning—but Scaling Requires Reliability Data

13 New Orleans Converts Waste Glass into Sand for a Local Coastal-Resilience Experiment

SCIENCE AND TECHNOLOGY FOR THE SDGs

14  Kaolin-Derived Zeolite and Chemistry-Informed AI Improve Dye-Wastewater Process Control

15  Whey-Protein Microcapsules Strengthen Gene-Silencing Treatment Against a Crop Fungus

01	GLOBAL GOVERNANCE AND SDG DELIVERY

HLPF 2026 Moves the SDG Targets to Integrated Delivery

DATE  15 July 2026   •   SDGs  6, 7, 9, 11 and 17EVIDENCE  Global policy review; 36 Voluntary National Reviews; ministerial declaration

The 2026 High-Level Political Forum closed in New York with a Ministerial Declaration and 36 Voluntary National Reviews, giving the world a sizeable but still self-reported evidence base on national SDG implementation. The Forum examined five goals—water and sanitation, clean energy, industry and infrastructure, sustainable cities, and partnerships—whose delivery systems overlap in practice. Water utilities need reliable power; renewable-energy expansion requires grids, finance and land-use decisions; and resilient cities depend on transport, drainage, housing and digital infrastructure. Treating these as independent sectors therefore understates both the risks of policy conflict and the gains from coordinated investment.¹²

The most important analytical shift was from commitments to implementation architecture. Countries with constrained fiscal space cannot simply add five separate sector programmes. They need common project pipelines, interoperable data, shared procurement and financing structures that evaluate multiple benefits. A solar-powered water scheme, for example, can advance SDGs 6, 7 and 13 while reducing recurrent energy costs. Conversely, infrastructure that expands access but increases emissions, displacement or debt can improve one indicator while weakening several others. Integrated planning is therefore not a slogan; it is a method for identifying these trade-offs before public money is committed.

The Forum’s limitations should also be explicit. VNRs are government-led and differ in statistical quality, time periods and willingness to disclose failure. Adoption of a declaration does not demonstrate financing or delivery. With roughly three and a half years remaining to 2030, the next useful test is whether governments publish costed, time-bound implementation plans and subnational data rather than further lists of aspirations. The HLPF has clarified the policy direction. Its credibility will now depend on measurable changes in budgets, service coverage, institutional coordination and outcomes for groups that national averages conceal.

02	GLOBAL GOVERNANCE AND SDG DELIVERY

Local2030 Puts Cities at the Centre of SDG Delivery

DATE  15 July 2026   •   SDGs  11 and 17, linked to 6, 7, 9 and 13EVIDENCE  International policy event; municipal implementation and localization framework

The Local2030 Coalition’s 15 July special event treated cities and local governments as delivery institutions rather than peripheral stakeholders. That emphasis is supported by the structure of the SDGs: UN-Habitat estimates that more than 65% of SDG targets—about 105 of 169—relate to local and regional government mandates. Municipalities make daily decisions on water, sanitation, solid waste, streets, housing, land use, public transport and disaster response. These functions create a practical chain from global targets to neighbourhood-level services.³⁴

Localization, however, requires more than adding SDG logos to municipal plans. A credible local system links targets to spatially disaggregated baselines, annual budgets, procurement, asset management and public reporting. Citywide averages can hide informal settlements where piped water, drainage, clean energy or safe transport remain unavailable. Geospatial information and digital public infrastructure can identify these gaps, but only when data standards, privacy safeguards and institutional responsibilities are clear. Voluntary Local Reviews can strengthen accountability, yet they are most useful when indicators are repeated over time and connected to expenditure and service-quality data.

The central constraint is fiscal and institutional capacity. Local governments are often responsible for high-cost services but have weak own-source revenue, limited borrowing authority and short project-preparation pipelines. International climate and development funds usually require feasibility studies, safeguards, audited accounts and measurable results—capabilities that smaller municipalities may lack. The Coalition’s agenda will therefore be judged by whether it helps cities convert plans into bankable, socially inclusive investments. A practical localization package would combine predictable intergovernmental transfers, technical support, neighbourhood data, participatory planning and access to blended finance.

03	GLOBAL GOVERNANCE AND SDG DELIVERY

 Healthy Diet Is Now a Macroeconomic Issue

DATE  14 July 2026   •   SDGs  2, 3, 12 and 17EVIDENCE  UN food-security event using latest available SOFI estimates

The HLPF special event on the high cost of healthy diets reframed nutrition as a purchasing-power problem as much as a production problem. The latest completed SOFI estimates available before the 2026 report showed that 2.60 billion people could not afford a healthy diet in 2024, down from 2.76 billion in 2019. Yet the global improvement concealed deterioration in poorer economies: the number unable to afford such a diet rose from 464 million to 545 million in low-income countries, and from 791 million to 869 million in lower-middle-income countries excluding India.⁵⁶

Inflation explains part of the divergence. Global food-price inflation peaked at 13.6% in January 2023, 5.1 percentage points above headline inflation; in low-income countries it reached about 30% in May 2023. Poor households devote a larger share of spending to food and have less capacity to substitute, save or absorb shocks. The policy implication is that expanding cereal supply alone will not guarantee diet quality. Prices of vegetables, fruit, pulses, eggs, dairy and other nutrient-dense foods, together with household income, transport and storage costs, determine whether a nominally available diet is actually affordable.

A serious response therefore combines agriculture, markets, public health and social protection. Climate-resilient production and cold chains can reduce volatility and post-harvest loss; school meals and targeted transfers can protect consumption; and competition, trade and market information can limit avoidable price transmission. Monitoring must also move beyond national averages to diet-cost data by region and income group. The event correctly elevated affordability, but future reporting should separate the number of people facing unaffordability from changes in dietary diversity, child nutrition and non-communicable disease. Lower diet cost is an intermediate outcome; improved nutrition is the development result.

04	JOBS, SKILLS AND SOCIAL PROTECTION

Paraguay’s Job Growth Is Substantial—Job Quality and Attribution Are the Next Questions

DATE  6 March 2026; reviewed 16 July 2026   •   SDGs  1, 5, 8 and 10EVIDENCE  Official labour-market statistics reported through the UN Global Accelerator

Paraguay reported 242,000 additional jobs since 2023, commonly rounded to nearly 250,000, against a government target of 500,000 by August 2028. That equals 48.4% of the target in roughly three years. More than six in ten of the new jobs were reported as formal, women accounted for 60%, and unemployment fell to 3.6%, its lowest level in more than 25 years.⁹ These details make the result stronger than a headline employment count: they suggest simultaneous progress in labour-force absorption, formalization and women’s participation.

The figures still require careful interpretation. If 60% of 242,000 jobs went to women, the implied number is about 145,200; if just over 60% were formal, at least roughly 145,000 jobs carried formal status. But the source does not provide a complete sectoral, wage, hours-worked or duration breakdown. A formal job may still be low-paid or temporary, and an increase in jobs can reflect broad economic recovery, demographic change or measurement effects as well as targeted policy. The evidence therefore establishes labour-market improvement, not the share directly caused by the Global Accelerator.

The policy architecture—tripartite engagement, skills, formalization and equal-opportunity mechanisms—is nevertheless relevant. Paraguay can strengthen the result by publishing quarterly flows into and out of employment, median earnings, social-security registration, youth and rural participation, and twelve-month retention. It should also report the baseline and methodology used to define an “additional” job. The programme is on pace numerically, but reaching the remaining 258,000 jobs by August 2028 will require maintaining creation while protecting quality. The development test is whether employment reduces household vulnerability and inequality, not simply whether a cumulative counter reaches 500,000. Publishing regional results would also reveal whether growth is concentrated in Asunción or spreading to poorer departments.

05	JOBS, SKILLS AND SOCIAL PROTECTION

Albania’s LEAP Programme Connects Training to Jobs

DATE  7 July 2026   •   SDGs  5, 8 and 10EVIDENCE  Programme monitoring with beneficiary case evidence

Albania’s Lifelong Empowerment and Protection programme reported vocational training for 617 people and employment for 120 women and young people in tourism, hospitality and care services. The observed placement-to-training ratio is 19.4%, although the groups are not necessarily identical and some trainees may still be completing courses. The programme also provided integrated support to 493 households and trained 173 care and social-service professionals. In the Lezhë region, 56 women and men received certified instruction combined with soft skills, digital skills, career counselling and employer-arranged traineeships.¹⁰

This delivery model addresses a common weakness in skills programmes: certificates are counted even when courses are disconnected from vacancies. LEAP instead combines technical preparation with workplace exposure and links to local businesses. The wider project context is important. Albania reports that about 25% of young people are not in employment, education or training, while social-protection expenditure is about 9.6% of GDP compared with a 20.5% European Union average. The programme therefore operates where labour-market exclusion and limited service capacity overlap.¹¹

The available evidence is encouraging but not an impact evaluation. There is no control group, counterfactual employment rate, wage comparison or published twelve-month retention figure. Stories of women moving from informal or unpaid work into employment explain the mechanism, but they cannot establish average impact. Future reporting should separate enrolment, completion, placement, formal contracts, earnings and retention; show results by sex, age, disability and social-assistance status; and compare the cost per sustained job with alternative interventions. LEAP’s strongest lesson is operational: training is more credible when it is designed backward from actual labour demand and combined with social support. Its next challenge is proving that placements are secure, adequately paid and durable.

06	JOBS, SKILLS AND SOCIAL PROTECTION

Uzbekistan’s Social-Insurance to Reach 5.8 Million More People

DATE  16 July 2026   •   SDGs  1, 3, 5, 8 and 10EVIDENCE  Adopted legal reform with projected additional coverage

Uzbekistan’s new social-insurance framework is expected to extend maternity and sickness benefits to 5.8 million additional people and establish a dedicated Social Insurance Fund.¹² The scale is material: coverage at this level could protect millions of workers and families against income loss during illness, childbirth and temporary inability to work. It also links decent work to public health, because households without income replacement often delay care, borrow at high cost or reduce food and education spending.

The 5.8 million figure must be described correctly. It is projected additional statutory coverage, not a count of people already registered, claiming or receiving benefits. Legal entitlement becomes social protection only when administrative systems can identify eligible workers, collect contributions where required, process claims and make adequate payments on time. Informal workers, migrants, people in remote regions and workers with irregular earnings are especially vulnerable to exclusion if registration is employer-dependent or digitally complex.

A meaningful implementation dashboard should therefore report active registrations, contribution density, claims submitted and approved, processing time, payment adequacy, appeals and complaints. Data should be disaggregated by sex, employment status, region and income level. Maternity protection also requires attention to whether benefits encourage women’s formal employment or create incentives for discriminatory hiring. Fiscal sustainability matters as well: benefit promises must be backed by transparent actuarial assumptions and reliable financing. Uzbekistan has completed an important policy step by defining wider rights and institutional responsibility. The development outcome will be demonstrated when covered workers actually receive predictable income protection and when illness or maternity causes fewer households to fall into poverty. Until those results are reported, 5.8 million remains a quantified ambition rather than a verified beneficiary count.

07	JOBS, SKILLS AND SOCIAL PROTECTION

Malawi’s Compensation Reform to Protect 4.2 Million Workers

DATE  16 July 2026   •   SDGs  1, 3, 8 and 10EVIDENCE  Validated draft legislation and social-insurance design

Malawi has validated a new Workers’ Compensation Bill and social-insurance model that could extend protection to more than 4.2 million workers.¹³ The proposed scale is significant in a labour market where many people work informally or without effective insurance against injury, disability and lost earnings. A workplace accident can generate three simultaneous shocks: treatment costs, loss of the injured worker’s income and unpaid care burdens within the household. Compensation can therefore advance health, poverty reduction and decent work together.

The central analytical distinction is between potential legal coverage and delivered compensation. At the reporting point, the measure was still a bill. Its eventual reach will depend on enactment, regulations, employer registration, financing, inspection, accessible claims and enforcement. If procedures require extensive documentation or travel, workers in agriculture, domestic work and small enterprises may remain outside the system despite being legally eligible. Benefit adequacy also matters: a low or delayed payment may not prevent debt or asset sales.

Implementation should be evaluated through the proportion of the labour force registered, claims per 100,000 workers, approval and rejection rates, average processing time, replacement rate relative to earnings, rehabilitation and return-to-work outcomes, and appeals resolved. Malawi’s associated work on recognition of prior learning and support for women’s livelihoods broadens the reform agenda, but those components also need quantified employment and income results. The 4.2 million estimate is valuable because it defines the intended population; it should not be reported as people already protected in practice. The reform becomes transformative only when injured workers can obtain timely, predictable and affordable redress.

08	JOBS, SKILLS AND SOCIAL PROTECTION

Cambodia’s Skills Programme Reached 271,000 Enrolments

DATE  Data through April 2026   •   SDGs  1, 4, 8 and 10EVIDENCE  Administrative enrolment and earlier graduate-employment data

Cambodia’s programme offering free technical and vocational education with allowances to 1.5 million young people from poor and vulnerable families had enrolled about 271,000 participants by April 2026. Of these, approximately 229,000 were trained through public TVET institutions and 42,000 through the Skills Development Fund. Enrolment therefore stood at 18.1% of the national target.¹⁴ The allowance component is analytically important because poor youth often cannot attend nominally free courses when transport, meals, equipment and forgone earnings remain unaffordable.

An earlier reporting snapshot provides partial outcome evidence. By late 2025, nearly 150,000 participants had enrolled; 79,285 had graduated and 51,591 were reported employed. On those figures, employment equalled about 65.1% of graduates and 34.4% of enrolments at that time.¹⁵ These ratios should not be treated as final programme impact because cohorts differ, some students remain in training, and the source does not establish employment duration, wages, formality or whether jobs were additional.

The programme’s integrated design—social assistance, vocational training and labour-market pathways—directly addresses participation barriers. Its risk is scale without matching demand: rapid enrolment can produce oversupply in popular trades or low completion where course quality, local vacancies and counselling are weak. Future reporting should publish cohort-based completion, placement within six months, twelve-month retention, earnings change, employer satisfaction and outcomes for women, rural youth and persons with disabilities. Cambodia has moved beyond a target-only story because it now reports enrolment and employment. The next test is whether the expansion preserves quality and converts public spending into sustained, productive work rather than short-term placements or certificates. Tracking cost per graduate and cost per sustained job would make value-for-money comparisons possible.

09	COMMUNITY-LED ADAPTATION AND CONSERVATION

Atacama Communities Turn Fog into Water

DATE  17 July 2026   •   SDGs  6, 11, 13 and 15EVIDENCE  Field pilots and site-specific collection measurements

Communities and researchers in Chile’s Atacama are using vertical mesh collectors to condense water droplets from coastal fog. Reported yields range widely—from about 0.2 to 4.9 litres per square metre per day—depending on site, season, wind and fog density.¹⁶ At that range, a 1,000-square-metre installation might theoretically collect 200 to 4,900 litres on a suitable day. The wide interval is the main analytical point: fog harvesting is highly location-specific and cannot be evaluated from collector area alone.

The technology has several advantages for isolated settlements. It requires little external energy, avoids further groundwater abstraction and can be installed modularly near nurseries, revegetation sites or small farms. Pilot collectors reported in the story gathered roughly 7 to 10 litres each on a foggy day. That volume can support seedlings or selected household uses, but it is not evidence that a settlement’s full drinking-water demand has been met. Potability also requires testing, treatment and clean storage because airborne pollutants, bird contact and collection surfaces can affect quality.

A sound comparison should calculate litres delivered per year, capital and maintenance cost per cubic metre, downtime, water quality and the cost of alternatives such as tanker supply, wells, pipelines or desalination. Community management is crucial: mesh damage, blocked channels and contaminated tanks can quickly reduce performance. Fog harvesting is therefore best viewed as a diversified adaptation measure rather than a universal solution. Its strongest contribution is that it converts a recurrent local climatic feature into water without pumping a stressed aquifer. Where measured yields are reliable and uses are matched to volume and quality, it can reduce vulnerability and support ecosystem restoration; where fog is intermittent, it must remain one component of a broader water-security plan.

10	COMMUNITY-LED ADAPTATION AND CONSERVATION

Palawan’s Turtle Patrollers Convert Knowledge into Capacity

DATE  17 July 2026   •   SDGs  8, 13, 14 and 15EVIDENCE  Community monitoring and protected-nest operations

Residents of San Vicente, Palawan, are being trained and paid to locate olive ridley turtle nests, monitor beaches and reduce losses from stray dogs, poaching, artificial light, high tides and human disturbance.¹⁷ The approach uses a resource that external conservation teams cannot easily replicate: repeated local observation of beaches, tides, tracks and nesting behaviour. A resident patroller can cover the same coastline across seasons and respond quickly when a nest is exposed or hatchlings become disoriented.

The intervention creates two types of value. Operationally, more surveillance hours increase the chance that nests are found, marked and protected. Socially, modest paid work and direct participation can shift conservation from an externally imposed restriction to a locally owned service. This matters in communities dependent on fisheries and tourism, where conservation rules can otherwise be perceived as costs with few local benefits. The reported case of a patroller who began in 2023 and advanced in responsibility by 2026 also suggests a pathway for building community expertise rather than relying only on short projects.

Evidence of population recovery, however, requires longer and more systematic monitoring. Finding more nests may reflect better search effort rather than more nesting turtles. The programme should therefore publish patrol kilometres and hours, nests located, eggs protected, predation and inundation rates, hatching success, hatchlings reaching the sea and repeat nesting by tagged adults. Comparing protected and unprotected sites would strengthen attribution. Climate change and coastal development also create risks that patrols alone cannot solve, including beach erosion, rising tides and lighting from tourism facilities. Community patrollers are a credible front line, but their work should be integrated with zoning, dog control, lighting standards, habitat protection and scientific monitoring. The immediate impact is safer management of identified nests; regional turtle recovery remains a longer-term outcome to be demonstrated.

11	COMMUNITY-LED ADAPTATION AND CONSERVATION

Vadun Sacred Zones Protects Benin’s Mangroves

DATE  15 July 2026   •   SDGs  13, 14, 15 and 16EVIDENCE  Community governance and site-level conservation observations

Coastal communities in Benin are using Vodun sacred zones and the authority of Zangbéto guardians to restrict entry, tree cutting and extraction in selected mangrove areas. The mechanism responds to severe ecological pressure: Benin reportedly lost about 29% of its mangrove cover between 1995 and 2015.¹⁸ Mangroves provide nursery habitat for fish and shellfish, stabilize shorelines and store large amounts of carbon; their loss therefore affects livelihoods, disaster risk and climate mitigation simultaneously.

The intervention’s distinctive feature is enforcement legitimacy. Formal environmental rules may be weak where agencies lack staff, transport or day-to-day presence. Sacred-site rules are embedded in belief, local authority and social accountability, making violations visible and sanctions immediate. FAO has also documented community-led restoration in southern Benin, indicating that spiritual protection can complement replanting, livelihood alternatives and formal management rather than operate in isolation.¹⁹

The evidence should not be overstated. A national 29% decline does not quantify the effect of each sacred zone, and intact patches may differ from unprotected areas in access, hydrology or settlement pressure. Stronger evaluation would map sanctuary boundaries, compare annual canopy change inside and outside protected zones, measure seedling survival and track fisheries indicators and household resource use. It should also examine who defines sacred rules and whether women, fishers and marginalized groups participate in decisions. Even with these limitations, the case offers a practical governance lesson: conservation can become lower-cost and more durable when it works through institutions that communities already recognize. Vodun protection is not a substitute for land rights, pollution control or national enforcement, but it can maintain ecological refuges and strengthen local stewardship while those broader systems develop. Monitoring should also ensure that conservation restrictions do not shift extraction pressure into neighbouring, less-protected wetlands.

12	COMMUNITY-LED ADAPTATION AND CONSERVATION

‘Bear-dar’ Demonstrates Successful Early Warning

DATE  11 June 2026; reported during the review period   •   SDGs  3, 9, 11, 13 and 15EVIDENCE  First confirmed operational detection involving three polar bears

An artificial-intelligence-assisted radar system known as Bear-dar recorded its first confirmed detection at the Eureka weather station in the Canadian High Arctic: a mother polar bear and two cubs were identified before reaching staff facilities.²⁰ The system analyses radar signals to distinguish likely bears from other moving objects, then sends an alert for human verification. Staff monitored the animals and used vehicles to guide them back toward sea ice without injury to people or bears.

The measurable evidence is narrow but real: one managed encounter involving three animals. Early warning can reduce surprise, panic and defensive shooting, particularly at remote stations where visibility, darkness and weather limit human observation. Radar can also operate continuously and at distances beyond a person’s immediate field of view. The adaptation relevance is clear as declining sea ice can keep bears on land longer and increase encounters near settlements, camps and waste-storage areas.

One successful event does not establish system-wide effectiveness. Deployment decisions require sensitivity and specificity, false-alarm rates, detection distance under different weather and terrain, uptime, maintenance needs and cost per protected site. Algorithms trained in one landscape may perform differently around buildings, moving vehicles or uneven ground. Community acceptance and alert protocols matter as much as detection: an alarm must lead to safe, culturally appropriate action. Bear-dar should therefore complement waste management, patrols, education and Indigenous knowledge rather than replace them. A multi-site trial with standardized incident records would show whether the technology consistently reduces dangerous proximity and lethal responses. For now, the system has moved from concept to operational proof, but not yet to demonstrated population- or community-level impact.

13	COMMUNITY-LED ADAPTATION AND CONSERVATION

New Orleans Converts Waste Glass into Sand for a Local Coastal-Resilience Experiment

DATE  16 July 2026   •   SDGs  11, 12, 13, 14 and 15EVIDENCE  Operating circular-economy system and early restoration observations

New Orleans-based Glass Half Full collects discarded bottles from residents and hospitality businesses, cleans and crushes them into sand and aggregate, and supplies the material for sandbags, construction uses and coastal-restoration trials.²¹ The model connects two local problems: glass that is difficult to process through conventional recycling and Louisiana’s continuing wetland and shoreline loss. A related project reports about 100 tonnes of recycled sand and plans for a 570-foot berm in Bayou Bienvenue.²²

The circular-economy logic is attractive. Locally recovered glass can reduce landfill disposal and partially substitute for mined sand, while public collection creates visible participation in resilience work. Early restoration sites reportedly show returning vegetation, retained sediment and wildlife use. These observations indicate ecological function may be developing, but they do not yet quantify restored acreage, elevation gain, storm protection or long-term survival. Coastal engineering also requires careful control of grain size, contaminants, transport, placement and effects on soils and aquatic life.

Impact assessment should track tonnes collected, processing yield, energy and transport emissions, cost per tonne of usable aggregate, area treated, sediment retention, vegetation survival and performance after major storms. The correct comparison is not simply glass versus landfill; it is recycled glass versus alternative restoration materials and waste-management systems across the full life cycle. The initiative’s strongest verified achievement is organizational: it has built a functioning local chain from household and business waste to a usable material. Its coastal contribution remains experimental and volume-constrained relative to Louisiana’s landscape-scale losses. Even so, it demonstrates how community participation, materials engineering and restoration finance can be combined.

14	SCIENCE AND TECHNOLOGY FOR THE SDGs

Kaolin-Derived Zeolite and Chemistry-Informed AI Improve Dye-Wastewater Process Control

DATE  16 July 2026   •   SDGs  6, 9, 12 and 14EVIDENCE  Laboratory adsorption tests and continuous-column prototype with external validation

Researchers converted kaolin clay into zeolite 4A and tested it for removal of dye from wastewater, then trained a chemistry-informed machine-learning model to predict adsorption performance. In the continuous-column experiment, one gram of zeolite treated ten litres of dye effluent and achieved an adsorption capacity of 56.6 milligrams per gram.²⁴ The team used 200 experimental observations for model development and 63 additional observations for external testing.

The model reported an R² of 0.9993 and root-mean-square error of 0.0493, compared with 0.0938 for an unconstrained two-stage Gaussian-process model—about a 47% reduction in prediction error. Its uncertainty intervals covered 96.8% of observations. The methodological contribution is important: mass balance, kinetics, saturation and stoichiometric rules were embedded so that predictions could not fit the data while violating known chemistry. Such constraints can improve trust when operators use models to select adsorbent quantities or operating conditions.

The environmental claim must remain bounded. The study addressed dye effluent under controlled conditions; it did not demonstrate safe drinking water, full industrial treatment or municipal-scale operation. Real wastewater contains competing ions, variable pH, suspended solids and multiple organic compounds that can reduce adsorption. Scale-up also requires evidence on hydraulic performance, zeolite regeneration, energy, cost, disposal of pollutant-loaded material and leaching risk. R² alone is not a measure of environmental benefit, especially when the experimental domain is narrow. The study nevertheless offers a credible integration of abundant mineral feedstock, materials science, column engineering and interpretable AI. The next stage should compare lifecycle cost and removal performance against established adsorbents using complex effluent over repeated regeneration cycles.

15	SCIENCE AND TECHNOLOGY FOR THE SDGs

Whey-Protein Microcapsules Strengthen Gene-Silencing Treatment Against a Crop Fungus

DATE  17 July 2026   •   SDGs  2, 3, 9, 12 and 15EVIDENCE  Laboratory assays and tomato bioassay; early-stage crop-protection research

Researchers tested short synthetic DNA molecules designed to suppress two genes associated with disease development in Verticillium dahliae, the soil-borne fungus that causes Verticillium wilt. To improve delivery, they compared chemical methylation with encapsulation in whey-protein-concentrate microcapsules measuring roughly 0.20 to 0.36 micrometres.²⁵ Both formulations reduced target-gene activity, but the encapsulated treatment produced significantly stronger suppression than methylated DNA, with P<.05, and the greatest reduction in fungal colony growth.

Tomato bioassays also showed lower disease severity. The whey-protein-encapsulated formulation produced a significantly lower area under the disease-progress curve than the methylated treatment and untreated pathogen control, with P<.001. Whey protein without the DNA treatment did not reduce disease, supporting the interpretation that the effect came from the gene-silencing cargo rather than the capsule material alone.

A targeted nucleic-acid treatment could reduce reliance on broad-spectrum fungicides and associated residues or non-target effects. Whey protein is a food-industry material and may offer a biodegradable carrier. However, the evidence remains pre-commercial. The researchers could not definitively confirm intracellular uptake, and the work was not a multi-location field trial. Soil conditions, ultraviolet exposure, rainfall, microbial degradation and plant growth stage may alter performance. Before agricultural use, studies must establish field efficacy, persistence, dose, production cost, non-target effects, environmental fate and regulatory acceptability. The results demonstrate an improved delivery strategy under laboratory and bioassay conditions; they do not yet establish practical disease control at farm scale. Comparative trials should include standard fungicides, untreated infection and integrated disease-management practices, with measurements of marketable yield .

METHOD AND SOURCE NOTE

The newsletter was prepared from the user-supplied compilation, then checked against original institutional reporting, international journalism and peer-reviewed research. Arithmetic ratios are calculated from published values and rounded to one decimal place where appropriate.

Quotations are intentionally brief. The analysis avoids implying causality where only programme monitoring, administrative data, policy announcements or early-stage experiments are available. Dates and titles were corrected where the original compilation could create a misleading chronology, including Brazil’s post-COP30 municipal climate work.

DEVELOPMENT CONNECTSFor a Better Tomorrow