Digital Platforms for Urban Tree Inventories and Green Infrastructure Management: A Multi-Criteria Assessment of Functional Maturity, Data Portability, and Tree-Safety Workflows

Abstract

Urban tree inventories increasingly support municipal green infrastructure management, but software selection can constrain data reuse, field workflows, environmental reporting, and tree-safety documentation. This study evaluated 29 anonymized digital platforms using an application-specific multi-criteria framework covering Management and Use, Data Storage and Export, Functionality and Features, Cost and Licensing, Environmental and Scientific Approach, and Tree Safety. Each platform was assessed once during the March–December 2025 study period using the functionality and evidence available at the time of assessment. Basic inventory, mapping, and administrative functions were generally more mature than environmental and scientific functions, data portability, and integrated tree-safety workflows. GIS-based inventory functions were widely available, whereas export and migration capabilities, ecosystem service support, advanced 3D/LiDAR integration, longitudinal tree-safety records, and reassessment alerts were less consistently implemented. Sensitivity analyses using alternative category weights produced similar aggregate conclusions, while the exploratory evidence-source comparison indicated that scores tended to be higher among platforms assessed with a live demonstration, although evidence-source bias could not be excluded. The findings highlight the importance of defining operational, data governance, environmental, and tree-safety requirements before selecting a platform and undertaking large-scale urban tree inventory programs.

1. Introduction

Urban trees are a core component of urban green infrastructure, but their long-term management depends on more than periodic field inspections. Municipal inventories require spatially explicit, updatable records connecting tree location, identity, dimensions, condition, maintenance history, and risk-related observations [1,2,3]. Digital inventory systems determine how these observations are structured, retrieved, compared, and reused over time.

The existence of an inventory does not, by itself, ensure strategic use. Keller and Konijnendijk compared inventories in six North American and Northern European cities and identified differences in updating and management integration [4]. In a later Swedish survey, 52.8% of responding municipalities had an inventory. Among these inventories, 54.1% were under development and 27.1% were outdated [5]. These studies distinguish inventory creation from the organizational capacity to maintain and use the resulting information.

Research has examined online mapping and image-assisted urban tree surveys [6,7], while recent work has considered digital tools across the inventory, assessment, planning, and monitoring stages of urban forestry [8]. Reviews of smart urban forestry and forest information requirements also connect digital data with management and environmental analysis [9,10]. These contributions establish the relevance of digital tools but do not imply that every inventory platform offers the same depth of implementation or integration.

Software selection is therefore a multi-criteria decision problem. General multi-criteria methods make criteria, scoring, and weighting choices explicit [11], while enterprise software selection research emphasizes translating user requirements into assessable functions [12]. For municipal inventories, this reasoning directs attention beyond purchase price or the presence of a map to usability, export, storage, reporting, licensing, and continued access to records.

Tree-safety work introduces further information requirements. Visual tree assessment (VTA) may be complemented by instrument-based techniques, including resistance drilling, sonic tomography, and pulling tests, when internal defects or structural stability require investigation [13,14]. For this application, the ability to attach an external report should be distinguished from structured assessment records, diagnostic data handling, intervention history, and reassessment scheduling. The present study evaluates software support for these functions, not the diagnostic accuracy of the instruments or the validity of individual tree risk decisions.

The relevant evidence addresses different units of analysis: inventory methods [2], municipal inventory practices [4,5], digital tool uses reported in research [8,9], and general software selection requirements [12]. These studies provide context rather than a directly comparable assessment of the six functional domains examined here. The contribution of this study is an application-specific synthesis of established multi-criteria principles, not a new generic decision-making algorithm. It links inventory operations, data portability, environmental functions, and tree-safety workflows within a common assessment.

The remaining question is how these functions coexist within platforms intended for practical urban tree management. We ask which core functions are widely documented, which advanced functions are less consistently implemented, and how these patterns can inform requirements before large-scale field collection. The aim is to assess functional implementation across relevant digital platforms and develop a cautious requirements checklist for municipal green infrastructure management without ranking identifiable products.

2. Materials and Methods

2.1. Platform Identification and Selection

Candidate platforms were identified iteratively through Google web searches, scientific articles, technical and professional publications, conference material, and recommendations from colleagues and practitioners. These sources were used to locate tree-specific inventory, mapping, or management systems, rather than to assemble a systematic review of publications.

Candidates were screened for tree-specific functionality relevant to inventory or management workflows. Generic geographic information system (GIS) products without such functions were excluded. Public documentation and provider communication were then considered to determine whether sufficient information was available for assessment. Candidates with insufficient evidence were excluded. This process resulted in 29 platforms. The distinction between tree-specific inventory systems and generic spatial tools was informed by inventory method and parameter research [1,2,3].

The identification strategy was exploratory and purposive, with no preregistered systematic search protocol. A complete search-by-search audit trail sufficient to reproduce exact historical query strings and counts of all initially identified, duplicate, or excluded candidates is not available for this report. Accordingly, no retrospective PRISMA flow counts are inferred from the final sample size. The results describe the selected platforms, not a census or probability sample of the software market.

2.2. Evaluation Period and Evidence Sources

Each platform was evaluated once during March–December 2025, using the functionality and evidence available at the time of assessment. The platforms were not all reassessed at a common final date. Live demonstrations informed the evaluation of 11 platforms; online sessions typically lasted approximately 30–45 min and covered mapping, data entry, reporting, user management, storage, export, mobile functionality, and tree-safety functions. The other 18 platforms were evaluated without a live demonstration using available public or provider-supplied information.

Live demonstration status recorded whether the research team received a direct presentation and explanation of a platform during the assessment, irrespective of whether a demonstration service was publicly advertised. Sessions included both routine commercial demonstrations and presentations arranged through research-related or professional contacts. The grouping therefore describes the evidence obtained during this study, rather than whether providers routinely offered a commercial demonstration service.

A predefined Excel-based matrix was completed during demonstrations. Functions shown directly were assessed from observed functionality. Described but undemonstrated functions were assessed using the supporting documentation and explanations available. Other evidence included official technical documentation, product descriptions, disclosed pricing, and provider correspondence. A function was not assumed to be present solely because it might plausibly be offered. This was an assessment of documented capabilities rather than an exhaustive test of every platform function.

Two researchers (D.H. and D.M.) participated throughout the assessment and discussed the interpretation of criteria and ambiguous cases. D.H. assigned the final sub-criterion scores for the complete dataset. Other participants attended individual demonstrations but did not independently score the complete sample. The common rubric and documented evidence provided a structured basis for judgment but did not remove the need for interpretation. No independent parallel rating dataset or formal inter-rater reliability estimate was produced.

2.3. Evaluation Categories and Sub-Criteria

Forty sub-criteria were organized into six author-developed functional domains: Management and Use (5 sub-criteria), Data Storage and Export (4), Functionality and Features (16), Cost and Licensing (5), Environmental and Scientific Approach (3), and Tree Safety (7). Inventory parameter research informed the link between data fields and management purposes [3,15]. General software selection and multi-criteria principles informed the use of explicit requirements [11,12]. The environmental information and tree assessment literature informed the relevance of scientific and safety-related functions [10,13,14]. This synthesis does not imply that an existing publication validated the same six-domain taxonomy.

The domains separate day-to-day access and usability, persistence and transfer of records, operational functions, commercial arrangements, ecological interpretation, and safety workflows. Some functions support more than one management purpose, but each sub-criterion contributes to only one category mean. Table 1 specifies the assessed scope and distinguishes the original numerical anchors from the judgment used between them.

Table 1. Evaluation categories, sub-criteria, and scoring anchors. Panel A lists the 40 sub-criteria assessed in 29 anonymized platforms, grouped into six functional domains. Definitions describe the scope of the assessment, not a claim that every listed operation was independently tested in every platform. Panel B preserves the original 0–10 scoring anchors and explains intermediate judgments. Abbreviations: GIS, geographic information system; VTA, visual tree assessment; AI, artificial intelligence; API, application programming interface; EU, European Union; GDPR, General Data Protection Regulation; LiDAR, light detection and ranging.

(A) Functional Scope and Operational Relevance
Category	Sub-criterion	Operational Definition	Operational Relevance
Management and Use	Reply within 1 day	Vendor responsiveness to the initial contact or request for information within one working day.	Describes initial responsiveness; not a longitudinal test of support quality.
Management and Use	User-friendly/cross-platform access	Ease of use and access across relevant platforms such as web, desktop and mobile environments.	Supports daily use by municipal staff, consultants and field crews.
Management and Use	Customizable tree forms	Capacity to adapt data entry fields to different inventory, inspection and management protocols.	Allows local VTA protocols, municipal requirements and future data fields to be added.
Management and Use	Unlimited users with permissions	User-number arrangements together with differentiated viewing and editing permissions; the combined criterion covers both scalability and access control.	Important for large teams, supervisors, contractors and quality control workflows.
Management and Use	Public tiered access	Controlled access for citizens, external partners or limited users with appropriate viewing/reporting permissions.	Supports public participation while protecting core database integrity.
Data Storage and Export	Sufficient online storage	Availability of adequate online or cloud storage for tree records, photographs and attachments.	Prevents storage limits from reducing the long-term usefulness of the inventory.
Data Storage and Export	Data storage expansion	Capacity to increase storage volume as the inventory grows.	Important when thousands of trees and repeated inspections are recorded over time.
Data Storage and Export	Documented hosting/data protection provisions	Documented hosting reliability and data protection arrangements relevant to EU public sector use; not a legal compliance certification.	Identifies matters requiring contractual and institutional verification before deployment.
Data Storage and Export	Data export/migration	Capacity to export or migrate tree records in usable formats.	Reduces vendor lock-in and protects the investment in field data collection.
Functionality and Features	Attachment support for external files	Ability to attach photographs, reports, diagnostic files or other documentation to tree records.	Supports evidence-based documentation and long-term traceability.
Functionality and Features	AI-driven insights	Availability of artificial intelligence-based interpretation or automated analytical support.	Represents advanced decision support capacity rather than simple record keeping.
Functionality and Features	Predictive maintenance tools	Tools that anticipate maintenance needs or future tree-related interventions.	Supports preventive management and resource planning.
Functionality and Features	API integration	Ability to connect with external databases, GIS systems or other digital infrastructure.	Improves interoperability and long-term data reuse.
Functionality and Features	Add-on modules	Availability of additional modules for specialized functions.	Shows whether the platform can expand beyond basic inventory use.
Functionality and Features	Mapping/GIS capabilities	Mapping individual trees and displaying spatial inventory information; a map or 3D-looking basemap alone is not analytical handling of tree-level 3D data.	Forms the spatial foundation of urban tree management.
Functionality and Features	Mobile app support	Availability of mobile tools for field data collection.	Supports direct recording during field surveys.
Functionality and Features	Offline mode	Ability to collect or edit data without an active internet connection.	Important for field crews working in areas with poor connectivity.
Functionality and Features	Local climate data integration	Ability to integrate climate or weather-related data.	Supports environmental interpretation and climate resilience analysis.
Functionality and Features	Automated reporting	Ability to generate reports automatically from stored data.	Reduces manual work and supports standardized reporting.
Functionality and Features	Customizable reports	Ability to adapt reports to local or institutional needs.	Supports professional, municipal and scientific reporting requirements.
Functionality and Features	Calculation/statistical functions	Internal calculations, summaries or basic analytical functions.	Supports quick interpretation without external processing.
Functionality and Features	Task assignment	Ability to assign and track maintenance or inspection tasks.	Supports operational management and work-order workflows.
Functionality and Features	Tree value calculation	Tools for estimating tree value or related economic indicators.	Supports valuation, compensation or benefit assessment.
Functionality and Features	Technical support	Support arrangements and assistance evidenced during the assessment; long-term service performance was not independently monitored.	Affects implementation, troubleshooting and long-term adoption.
Functionality and Features	Backup and recovery	Availability of backup and recovery functions.	Protects the inventory against data loss.
Cost and Licensing	Free trial/demo	Availability of trial access or demonstration before purchase.	Allows users to verify functions before procurement.
Cost and Licensing	Flexible pricing	Availability of flexible pricing models.	Supports adoption by different municipality sizes and project budgets.
Cost and Licensing	No hidden fees	Clarity of disclosed charges and additional fees in the available pricing information; not a guarantee about all future costs.	Reduces procurement risk and unexpected long-term costs.
Cost and Licensing	License flexibility	Flexibility of subscription, licensing or user arrangements.	Affects scalability and long-term affordability.
Cost and Licensing	Data retention after non-payment	Documented retention, access, or export arrangements when a subscription ends; not an empirical test of every future termination scenario.	Protects the value of collected field data and reduces vendor lock-in.
Environmental and Scientific Approach	Environmental impact/ecosystem service tools	Tools related to ecosystem services or environmental benefit assessment.	Supports environmental reporting and green infrastructure planning.
Environmental and Scientific Approach	Tree health monitoring	Structured support for monitoring tree condition over time.	Supports long-term scientific and management interpretation.
Environmental and Scientific Approach	Biodiversity analysis	Tools related to species diversity or biodiversity indicators.	Supports ecological interpretation and biodiversity reporting.
Tree Safety	VTA integration	Structured support for visual tree assessment records.	Supports tree risk assessment and repeatable safety documentation.
Tree Safety	Instrument data upload	Ability to link uploaded diagnostic outputs or external safety analyses to tree records; upload does not imply native processing.	Supports documentation of resistance-drilling, tomography, or pulling-test results.
Tree Safety	LiDAR/3D data support	Support for LiDAR, point clouds, or 3D models. Distinguish attached/external visualization, native data handling linked to tree records, and analytical use such as measurement or comparison.	Distinguishes display from analytical support. No universal file format or task-based benchmark was applied.
Tree Safety	Automated risk report generation	Ability to generate risk-related reports automatically.	Supports standardized risk communication and documentation.
Tree Safety	Visual report customization	Ability to customize the visual appearance and content of safety reports.	Supports professional reporting and communication to decision-makers.
Tree Safety	Dynamic tree history management	Ability to store and compare repeated assessments or events over time.	Supports monitoring, reassessment and longitudinal interpretation.
Tree Safety	Reassessment alerts/reminders	Automatic reminders for scheduled reassessments or follow-up inspections.	Supports proactive risk management and monitoring.
(B) Original Anchors and Interpretation of Intermediate Scores
Score	Role	Interpretation
0	Anchor	Function absent under the scoring rubric. Lack of public information alone does not independently establish absence.
1–2	Intermediate judgment	Position between anchors 0 and 3. Separate standardized tests for these values were not specified.
3	Anchor	Function not currently operational but potentially developable, customizable, or activatable at additional cost.
4	Intermediate judgment	Position between anchors 3 and 5; below the reporting cut-off for at least partial implementation.
5	Anchor	Partial availability, dependence on additional configuration, a separate module, or a non-standard workflow.
6–8	Intermediate judgment	Available/usable functionality with differing limitations in completeness, usability, automation, and workflow integration. Distinctions within this range required evaluator judgment.
9	Near-complete anchor	Almost fully integrated and operational, with minor limitations.
10	Anchor	Fully integrated and directly operational without substantial additional development or technical intervention. Novelty alone was not the scoring anchor.

2.4. Scoring Scale and Workflow

The 0–10 scores are an ordinal, semi-quantitative representation of qualitative functional maturity. For each sub-criterion, available evidence was considered against the functional scope in Table 1, matched to the original anchor descriptions, and assigned a numerical score. The anchors were 0, 3, 5, and 10, with 9 describing near-complete implementation. Intermediate integers expressed judgments about completeness, usability, automation, and integration rather than measurements in equally spaced physical units. A score of 4 was an intermediate judgment between the non-operational/potential implementation anchor at 3 and the partial implementation anchor at 5.

Table 1 does not retrospectively introduce eleven separately validated performance classes. The exact distance between neighboring scores was not empirically calibrated. Applying the same rubric made the assessment structure explicit, but agreement between independent evaluators was not measured. The numerical aggregation described below is therefore a declared analytical convention for these assigned scores, not evidence that the underlying qualitative differences are equal in magnitude.

Descriptive reporting separates scores below 4, scores in the interval 4 ≤ s < 5, and scores of at least 5. These three intervals cover the full scale. The additional cut-offs ≥ 7 and ≥ 9 identify nested subsets of the ≥5 group. They do not form extra mutually exclusive classes. For integer sub-criterion scores, the middle interval contains only 4; for category and overall averages, it also contains non-integer scores between 4 and 5. Every observation remains in the denominator of 29. Labels such as partial, usable/strong, and near-full/full refer to the rubric, not independent certification of performance.

2.5. Data Analysis

For platform i and category k, the category score C_ik was the arithmetic mean of its m_k sub-criterion scores. The primary overall score was O_i = (C_i1 + C_i2 + C_i3 + C_i4 + C_i5 + C_i6)/6. Thus, every category received weight 1/6 despite having a different number of sub-criteria. Within a category, each sub-criterion received an effective overall weight of 1/(6m_k). This equal-category model was an exploratory baseline, not an empirical estimate of municipal preferences. Weighting and aggregation are explicit modeling choices consistent with general multi-criteria practice [11].

Means, sample standard deviations (SDs), counts, and percentages describe the 29 assessed platforms. Standard error (SE) was calculated as SD/√29 and retained alongside the means for consistency with the original reporting. These SE values do not quantify uncertainty from evaluator judgment, incomplete evidence, purposive selection, or software updates, and they are not design-based errors for estimates of the global market. Thresholds were applied to unrounded scores before display rounding. Category comparisons remained descriptive; no omnibus or pairwise category significance tests were performed. The domains summarize different constructed functions, and the study did not specify a population-level hypothesis of category equality.

The original scoring and descriptive tables were prepared in Microsoft Excel for Microsoft 365 (Microsoft Corporation, Redmond, WA, USA). During revision, numerical checks and the two additional analyses below were performed using Python 3.13.5, NumPy 2.3.5, and SciPy 1.17.0. The original sub-criterion scores and equal-weight baseline were retained. The additional analyses do not establish the reliability or interval-scale validity of the scoring rubric.

Cost categories were descriptive and were not based on universal monetary thresholds. Free/open access indicated that no direct license or subscription fee was identified for core platform use, although hosting, configuration, or technical labor could still incur costs. Low, medium, and high were relative categories based on the disclosed acquisition or subscription prices within the evaluated sample. Enterprise/service-based referred to customized organization-level packages that could combine software, hosting, implementation, training, or field services. Service-based/vendor-managed referred to cases in which the provider conducted or managed the inventory and supplied platform access, outputs, or reports rather than selling a standalone software product. Offers without sufficient price transparency were classified as undisclosed.

2.5.1. Weighting Sensitivity Analysis

Two illustrative scenarios recalculated each overall score from the unchanged category scores. One assigned 0.30 to Tree Safety and 0.14 to each remaining category. The other assigned 0.30 to Data Storage and Export and 0.14 to each other category. The weights sum to 1 and are illustrative priority scenarios, not preferences elicited from municipalities. We compared the sample mean, mean absolute within-platform score change, and counts crossing the reporting cut-offs. Only aggregate results are reported. No platform ranking or score profile is disclosed.

2.5.2. Evidence-Source Comparison

The overall score distributions of platforms with live demonstrations (n = 11) and those evaluated without a live demonstration (n = 18) were compared in one exploratory, two-sided Mann–Whitney U test. The asymptotic calculation used continuity correction; there were no tied overall scores. We report medians, first and third quartiles (linear interpolation), U for the demonstration group, the p-value, and the rank-biserial effect size r_rb = 2U/(11 × 18)−1. A positive r_rb indicates higher scores in the demonstration group. This compares score distributions, not necessarily medians alone. Since receipt of a live demonstration was not randomly assigned, the comparison cannot isolate the causal effect of evidence source or establish equivalence between groups.

2.6. Anonymization and Confidentiality

Platform and company names, product URLs, countries of origin, exact prices, and platform-level score profiles are not reported. Even coded profiles combined with distinctive characteristics could permit identification through public information. The study therefore reports aggregate functional patterns rather than a commercial ranking. This reporting choice preserves the intended anonymity but limits external verification of sample composition and replication of product-level scoring. It should not be interpreted as a guarantee that commercial neutrality follows automatically from anonymization.

The authors retain the original scoring matrix, evaluation records, demonstration material, provider correspondence, source documentation, and calculation sheets for verification. The report includes the rubric, category structure, aggregation method, and aggregate results but not a platform-level appendix. Access to non-identifying methodological information is subject to the restrictions stated in the Data Availability Statement.

3. Results

All counts and percentages refer to the purposive sample of 29 platforms. The following subsections follow the six categories in Table 1. They describe the assigned scores and documented capabilities, not population estimates or comparative product recommendations.

3.1. Overall Functional Maturity

The mean overall score was 4.30 ± 0.32 (mean ± SE), with SD = 1.74 and a range of 2.11–7.49. Fifteen platforms (51.7%) scored below 4, five (17.2%) scored from 4 to below 5, and nine (31.0%) scored at least 5. Three platforms (10.3%) reached an overall score of at least 7; none reached 9. Table 2 and Figure 1 summarize the overall and category-level distributions.

Table 2. Overall and category-level score summaries. Scores (s) range from 0 to 10; higher scores indicate stronger documented functional implementation. Across 29 platforms, mean ± standard error (SE, sample SD/√29) and sample standard deviation (SD) summarize the assigned scores. The columns < 4 (low scores), 4 ≤ s < 5 (between the low-score and partial implementation cut-offs), and ≥5 (at least partial implementation) exhaust the sample. The ≥7 (usable/strong) and ≥9 (near-full/full) columns are nested subsets of ≥5, not additional exclusive classes. Each percentage is 100n/29. These are rubric-based summaries, not population estimates or independent verification of product performance. Category and overall values are averages; exceeding a category cut-off does not mean that every constituent function reaches that level.

Item	Mean ± SE	SD	<4 n (%)	4 ≤ s < 5 n (%)	≥5 n (%)	≥7 n (%)	≥9 n (%)
Overall score	4.30 ± 0.32	1.74	15 (51.7%)	5 (17.2%)	9 (31.0%)	3 (10.3%)	0 (0.0%)
Management and Use	4.43 ± 0.33	1.80	12 (41.4%)	6 (20.7%)	11 (37.9%)	5 (17.2%)	0 (0.0%)
Data Storage and Export	4.50 ± 0.46	2.50	12 (41.4%)	3 (10.3%)	14 (48.3%)	6 (20.7%)	1 (3.4%)
Functionality and Features	4.22 ± 0.44	2.35	16 (55.2%)	2 (6.9%)	11 (37.9%)	4 (13.8%)	1 (3.4%)
Cost and Licensing	6.41 ± 0.47	2.53	4 (13.8%)	4 (13.8%)	21 (72.4%)	14 (48.3%)	7 (24.1%)
Environmental and Scientific Approach	2.60 ± 0.67	3.61	20 (69.0%)	1 (3.4%)	8 (27.6%)	5 (17.2%)	3 (10.3%)
Tree Safety	3.66 ± 0.60	3.22	16 (55.2%)	0 (0.0%)	13 (44.8%)	7 (24.1%)	0 (0.0%)

Figure 1. Mean scores for six functional categories across 29 anonymized urban tree inventory and management platforms. Each platform’s category score is the arithmetic mean of its constituent 0–10 sub-criterion scores. Bars show category means, and horizontal error bars show standard error (sample SD/√29). Higher values indicate stronger implementation under the assigned rubric. Error bars do not represent inter-rater agreement, temporal stability, or probability sampling uncertainty, and no category significance comparison is implied. SD: sample standard deviation.

Cost and Licensing had the highest category mean (6.41 ± 0.47), followed by Data Storage and Export (4.50 ± 0.46), Management and Use (4.43 ± 0.33), and Functionality and Features (4.22 ± 0.44). Tree Safety (3.66 ± 0.60) and Environmental and Scientific Approach (2.60 ± 0.67) had lower means. These are descriptive category differences.

3.2. Management and Use

User-friendly or cross-platform access was at least partially implemented in 25 platforms (86.2%) and reached the ≥7 cut-off in 19 (65.5%). Customizable forms reached these thresholds in 22 (75.9%) and 17 platforms (58.6%). Unlimited users with differentiated permissions were at least partially implemented in 14 platforms (48.3%) and usable/strong in eight (27.6%) (Table 3).

Table 3. Management, usability, and access control scores. Scores (s) range from 0 to 10; higher scores indicate stronger documented functional implementation. Across 29 platforms, mean ± standard error (SE, sample SD/√29) and sample standard deviation (SD) summarize the assigned scores. The columns < 4 (low scores), 4 ≤ s < 5 (between the low-score and partial implementation cut-offs), and ≥5 (at least partial implementation) exhaust the sample. The ≥7 (usable/strong) and ≥9 (near-full/full) columns are nested subsets of ≥5, not additional exclusive classes. Each percentage is 100n/29. These are rubric-based summaries, not population estimates or independent verification of product performance. Public tiered access concerns controlled public or external user access. Responsiveness refers only to initial contact, not long-term support quality.

Item	Mean ± SE	SD	<4 n (%)	4 ≤ s < 5 n (%)	≥5 n (%)	≥7 n (%)	≥9 n (%)
Reply within 1 day	2.48 ± 0.78	4.19	21 (72.4%)	0 (0.0%)	8 (27.6%)	7 (24.1%)	6 (20.7%)
User-friendly/cross-platform access	7.21 ± 0.44	2.40	3 (10.3%)	1 (3.4%)	25 (86.2%)	19 (65.5%)	12 (41.4%)
Customizable tree forms	6.24 ± 0.73	3.92	7 (24.1%)	0 (0.0%)	22 (75.9%)	17 (58.6%)	13 (44.8%)
Unlimited users with permissions	3.83 ± 0.67	3.63	13 (44.8%)	2 (6.9%)	14 (48.3%)	8 (27.6%)	2 (6.9%)
Public tiered access	2.41 ± 0.81	4.35	22 (75.9%)	0 (0.0%)	7 (24.1%)	7 (24.1%)	7 (24.1%)

Public tiered access received the maximum score in seven platforms (24.1%) and a score of zero in the remaining 22 (75.9%). No intermediate scores were assigned to this criterion. Provider responsiveness had a mean score of 2.48 ± 0.78. Eight platforms (27.6%) reached at least partial implementation under this criterion, and five (17.2%) received the maximum score.

3.3. Data Storage and Export

Sufficient online storage reached at least partial implementation in 25 platforms (86.2%) and usable/strong implementation in 14 (48.3%). Export or migration reached these thresholds in 17 (58.6%) and six platforms (20.7%). Storage expansion reached the ≥7 cut-off in nine platforms (31.0%), while documented hosting and data protection provisions reached it in eight (27.6%) (Table 4). The latter criterion reflects the evidence assessed, not an independent finding of legal compliance or non-compliance.

Table 4. Storage, hosting, and export scores. Scores (s) range from 0 to 10; higher scores indicate stronger documented functional implementation. Across 29 platforms, mean ± standard error (SE, sample SD/√29) and sample standard deviation (SD) summarize the assigned scores. The columns < 4 (low scores), 4 ≤ s < 5 (between the low-score and partial implementation cut-offs), and ≥5 (at least partial implementation) exhaust the sample. The ≥7 (usable/strong) and ≥9 (near-full/full) columns are nested subsets of ≥5, not additional exclusive classes. Each percentage is 100n/29. These are rubric-based summaries, not population estimates or independent verification of product performance. Hosting/data protection scores concern documented arrangements relevant to EU public sector use; they are not legal or cybersecurity certifications.

Item	Mean ± SE	SD	<4 n (%)	4 ≤ s < 5 n (%)	≥5 n (%)	≥7 n (%)	≥9 n (%)
Sufficient online storage	6.24 ± 0.55	2.98	4 (13.8%)	0 (0.0%)	25 (86.2%)	14 (48.3%)	7 (24.1%)
Data storage expansion	5.03 ± 0.70	3.77	9 (31.0%)	1 (3.4%)	19 (65.5%)	9 (31.0%)	7 (24.1%)
Documented hosting/data protection provisions	2.66 ± 0.82	4.41	21 (72.4%)	0 (0.0%)	8 (27.6%)	8 (27.6%)	7 (24.1%)
Data export/migration	4.07 ± 0.63	3.40	11 (37.9%)	1 (3.4%)	17 (58.6%)	6 (20.7%)	4 (13.8%)

3.4. Functionality and Features

Mapping/GIS capabilities were at least partially implemented in 25 platforms (86.2%) and usable/strong in 18 (62.1%). Mobile support reached the same cut-offs in 20 (69.0%) and 13 platforms (44.8%), and offline operation in 15 (51.7%) and 12 platforms (41.4%). Table 5 reports these functions within their original category rather than repeating them in a separate core functions table.

Table 5. Operational and analytical functionality scores. Scores (s) range from 0 to 10; higher scores indicate stronger documented functional implementation. Across 29 platforms, mean ± standard error (SE, sample SD/√29) and sample standard deviation (SD) summarize the assigned scores. The columns < 4 (low scores), 4 ≤ s < 5 (between the low-score and partial implementation cut-offs), and ≥5 (at least partial implementation) exhaust the sample. The ≥7 (usable/strong) and ≥9 (near-full/full) columns are nested subsets of ≥5, not additional exclusive classes. Each percentage is 100n/29. These are rubric-based summaries, not population estimates or independent verification of product performance. GIS: geographic information system; API: application programming interface; AI: artificial intelligence. The functions cover field entry, documentation, interoperability, work management, reporting, and analysis.

Item	Mean ± SE	SD	<4 n (%)	4 ≤ s < 5 n (%)	≥5 n (%)	≥7 n (%)	≥9 n (%)
Attachment support for external files	5.76 ± 0.64	3.43	8 (27.6%)	0 (0.0%)	21 (72.4%)	15 (51.7%)	7 (24.1%)
AI-driven insights	1.17 ± 0.48	2.56	24 (82.8%)	1 (3.4%)	4 (13.8%)	2 (6.9%)	1 (3.4%)
Predictive maintenance tools	2.90 ± 0.71	3.83	17 (58.6%)	1 (3.4%)	11 (37.9%)	6 (20.7%)	4 (13.8%)
API integration	4.62 ± 0.85	4.57	13 (44.8%)	1 (3.4%)	15 (51.7%)	12 (41.4%)	10 (34.5%)
Add-on modules	5.38 ± 0.77	4.13	10 (34.5%)	1 (3.4%)	18 (62.1%)	14 (48.3%)	8 (27.6%)
Mapping/GIS capabilities	7.28 ± 0.61	3.30	4 (13.8%)	0 (0.0%)	25 (86.2%)	18 (62.1%)	15 (51.7%)
Mobile app support	5.66 ± 0.76	4.08	8 (27.6%)	1 (3.4%)	20 (69.0%)	13 (44.8%)	10 (34.5%)
Offline mode	4.59 ± 0.84	4.55	13 (44.8%)	1 (3.4%)	15 (51.7%)	12 (41.4%)	10 (34.5%)
Local climate data integration	1.93 ± 0.68	3.67	22 (75.9%)	0 (0.0%)	7 (24.1%)	4 (13.8%)	4 (13.8%)
Automated reporting	4.24 ± 0.67	3.61	11 (37.9%)	0 (0.0%)	18 (62.1%)	10 (34.5%)	3 (10.3%)
Customizable reports	4.59 ± 0.73	3.91	11 (37.9%)	0 (0.0%)	18 (62.1%)	12 (41.4%)	6 (20.7%)
Calculation/statistical functions	3.14 ± 0.65	3.51	15 (51.7%)	1 (3.4%)	13 (44.8%)	4 (13.8%)	3 (10.3%)
Task assignment	4.41 ± 0.82	4.42	14 (48.3%)	0 (0.0%)	15 (51.7%)	11 (37.9%)	8 (27.6%)
Tree value calculation	2.41 ± 0.70	3.75	20 (69.0%)	1 (3.4%)	8 (27.6%)	5 (17.2%)	4 (13.8%)
Technical support	4.69 ± 0.77	4.12	11 (37.9%)	0 (0.0%)	18 (62.1%)	11 (37.9%)	8 (27.6%)
Backup and recovery	4.69 ± 0.82	4.39	12 (41.4%)	0 (0.0%)	17 (58.6%)	11 (37.9%)	9 (31.0%)

Attachment support reached the ≥5 and ≥7 cut-offs in 21 (72.4%) and 15 platforms (51.7%). Add-on modules reached them in 18 (62.1%) and 14 platforms (48.3%). Automated reporting and customizable reports were at least partially implemented in 18 platforms (62.1%) each but reached the ≥7 cut-off in 10 (34.5%) and 12 (41.4%), respectively.

AI-driven insights were at least partially implemented in four platforms (13.8%) and usable/strong in two (6.9%). Predictive maintenance tools reached these cut-offs in 11 (37.9%) and six platforms (20.7%). Local climate integration reached them in seven (24.1%) and four (13.8%), and tree value calculation in eight (27.6%) and five (17.2%).

3.5. Cost and Licensing

The sample included four free/open-access platforms (13.8%), four low-cost platforms (13.8%), six medium-cost platforms (20.7%), five high-cost platforms (17.2%), one enterprise/service-based platform (3.4%), two service-based/vendor-managed platforms (6.9%), and seven with undisclosed or insufficiently transparent pricing (24.1%). These are relative commercial categories, not universal monetary bands (Table 6A).

Table 6. Cost arrangements and licensing scores. Panel A classifies the 29 platforms by the relative commercial arrangements disclosed during assessment, with percentages calculated as 100n/29. Price bands have no universal monetary boundaries and do not measure total cost of ownership. Panel B reports 0–10 licensing scores. Scores (s) range from 0 to 10; higher scores indicate stronger documented functional implementation. Across 29 platforms, mean ± standard error (SE, sample SD/√29) and sample standard deviation (SD) summarize the assigned scores. The columns < 4 (low scores), 4 ≤ s < 5 (between the low-score and partial implementation cut-offs), and ≥5 (at least partial implementation) exhaust the sample. The ≥7 (usable/strong) and ≥9 (near-full/full) columns are nested subsets of ≥5, not additional exclusive classes. Each percentage is 100n/29. These are rubric-based summaries, not population estimates or independent verification of product performance.

(A) Broad Cost Categories
Cost Category	n	% of Sample	Interpretation
Free/open access	4	13.8%	No direct license or subscription fee identified for core platform use; hosting, configuration, or technical labor may still incur costs.
Low cost	4	13.8%	Lower disclosed acquisition or subscription cost relative to the evaluated paid platforms; no universal monetary cut-off was applied.
Medium cost	6	20.7%	Intermediate disclosed acquisition or subscription cost relative to the evaluated paid platforms; no universal monetary cut-off was applied.
High cost	5	17.2%	Higher disclosed acquisition or subscription cost relative to the evaluated paid platforms; no universal monetary cut-off was applied.
Enterprise/service-based	1	3.4%	Customized organization-level quotation that may bundle software, hosting, implementation, training, or field inventory services.
Service-based/vendor-managed	2	6.9%	The provider conducts or manages the inventory and supplies platform access, outputs, or reports rather than selling a standalone product.
Undisclosed	7	24.1%	Insufficiently transparent pricing information was available for classification beyond undisclosed status.
(B) Licensing-Related Functions
Item		Mean ± SE	SD	<4 n(%)	4 ≤ s < 5 n(%)	≥5 n(%)	≥7 n(%)	≥9 n(%)
Free trial/demo		6.55 ± 0.78	4.21	7 (24.1%)	1 (3.4%)	21 (72.4%)	17 (58.6%)	15 (51.7%)
Flexible pricing		6.34 ± 0.64	3.47	4 (13.8%)	1 (3.4%)	24 (82.8%)	13 (44.8%)	11 (37.9%)
No hidden fees		5.83 ± 0.74	3.98	7 (24.1%)	1 (3.4%)	21 (72.4%)	13 (44.8%)	11 (37.9%)
License flexibility		5.34 ± 0.73	3.95	8 (27.6%)	1 (3.4%)	20 (69.0%)	11 (37.9%)	9 (31.0%)
Data retention after non-payment		8.00 ± 0.57	3.06	2 (6.9%)	1 (3.4%)	26 (89.7%)	21 (72.4%)	18 (62.1%)

Data retention after non-payment was at least partially implemented in 26 platforms (89.7%) and usable/strong in 21 (72.4%). Trial or demonstration availability reached these cut-offs in 21 (72.4%) and 17 (58.6%) platforms. Flexible pricing reached them in 24 (82.8%) and 13 (44.8%), clarity of additional charges in 21 (72.4%) and 13 (44.8%), and license flexibility in 20 (69.0%) and 11 (37.9%) (Table 6B). Availability of a trial or demonstration is a scored feature, distinct from whether a demonstration was actually used in this study.

3.6. Environmental and Scientific Approach

Environmental impact or ecosystem service tools were at least partially implemented in 10 platforms (34.5%) and usable/strong in nine (31.0%). Tree health monitoring reached the same cut-offs in 10 (34.5%) and six platforms (20.7%). Biodiversity analysis reached them in seven (24.1%) and four (13.8%) (Table 7).

Table 7. Environmental and scientific function scores. Scores (s) range from 0 to 10; higher scores indicate stronger documented functional implementation. Across 29 platforms, mean ± standard error (SE, sample SD/√29) and sample standard deviation (SD) summarize the assigned scores. The columns < 4 (low scores), 4 ≤ s < 5 (between the low-score and partial implementation cut-offs), and ≥5 (at least partial implementation) exhaust the sample. The ≥7 (usable/strong) and ≥9 (near-full/full) columns are nested subsets of ≥5, not additional exclusive classes. Each percentage is 100n/29. These are rubric-based summaries, not population estimates or independent verification of product performance. These sub-criteria concern software support for environmental analysis and monitoring, not independent validation of ecological models or benefit estimates.

Item	Mean ± SE	SD	<4 n (%)	4 ≤ s < 5 n (%)	≥5 n (%)	≥7 n (%)	≥9 n (%)
Environmental impact/ecosystem service tools	3.14 ± 0.80	4.28	18 (62.1%)	1 (3.4%)	10 (34.5%)	9 (31.0%)	6 (20.7%)
Tree health monitoring	2.72 ± 0.75	4.02	19 (65.5%)	0 (0.0%)	10 (34.5%)	6 (20.7%)	5 (17.2%)
Biodiversity analysis	1.93 ± 0.68	3.67	22 (75.9%)	0 (0.0%)	7 (24.1%)	4 (13.8%)	4 (13.8%)

3.7. Tree Safety

Structured VTA integration reached the ≥5 and ≥7 cut-offs in 18 (62.1%) and 12 platforms (41.4%). Instrument data upload reached them in 17 (58.6%) and 11 (37.9%), and automated risk-report generation in 17 (58.6%) and nine (31.0%) (Table 8).

Table 8. Tree-safety and diagnostic data support scores. Scores (s) range from 0 to 10; higher scores indicate stronger documented functional implementation. Across 29 platforms, mean ± standard error (SE, sample SD/√29) and sample standard deviation (SD) summarize the assigned scores. The columns < 4 (low scores), 4 ≤ s < 5 (between the low-score and partial implementation cut-offs), and ≥5 (at least partial implementation) exhaust the sample. The ≥7 (usable/strong) and ≥9 (near-full/full) columns are nested subsets of ≥5, not additional exclusive classes. Each percentage is 100n/29. These are rubric-based summaries, not population estimates or independent verification of product performance. VTA: visual tree assessment; LiDAR: light detection and ranging. Upload or display support is distinct from native analytical processing and does not establish the validity of a tree risk decision.

Item	Mean ± SE	SD	<4 n (%)	4 ≤ s < 5 n (%)	≥5 n (%)	≥7 n (%)	≥9 n (%)
VTA integration	4.93 ± 0.75	4.05	10 (34.5%)	1 (3.4%)	18 (62.1%)	12 (41.4%)	8 (27.6%)
Instrument data upload	4.69 ± 0.77	4.13	11 (37.9%)	1 (3.4%)	17 (58.6%)	11 (37.9%)	8 (27.6%)
LiDAR/3D data support	2.83 ± 0.73	3.92	18 (62.1%)	0 (0.0%)	11 (37.9%)	6 (20.7%)	5 (17.2%)
Automated risk report generation	4.24 ± 0.74	3.98	12 (41.4%)	0 (0.0%)	17 (58.6%)	9 (31.0%)	6 (20.7%)
Visual report customization	4.00 ± 0.72	3.86	13 (44.8%)	0 (0.0%)	16 (55.2%)	9 (31.0%)	5 (17.2%)
Dynamic tree history management	3.45 ± 0.73	3.92	15 (51.7%)	0 (0.0%)	14 (48.3%)	7 (24.1%)	5 (17.2%)
Reassessment alerts/reminders	1.48 ± 0.63	3.41	24 (82.8%)	0 (0.0%)	5 (17.2%)	4 (13.8%)	3 (10.3%)

LiDAR/three-dimensional support was at least partially implemented in 11 platforms (37.9%) and usable/strong in six (20.7%). Dynamic tree history reached the same cut-offs in 14 (48.3%) and seven (24.1%), while reassessment alerts reached them in five (17.2%) and four (13.8%). Visual report customization was at least partially implemented in 16 platforms (55.2%) and usable/strong in nine (31.0%).

3.8. Sensitivity and Evidence-Source Checks

Increasing the Tree Safety weight to 0.30 changed the mean overall score from 4.30 ± 0.32 to 4.20 ± 0.35. Increasing the Data Storage and Export weight to 0.30 produced 4.34 ± 0.33. Mean absolute within-platform changes were 0.29 and 0.23 score units, respectively. The number scoring at least 5 changed from nine (31.0%) to 10 (34.5%) in the Tree Safety scenario and remained at nine in the export-priority scenario. In both scenarios, 15 platforms (51.7%) remained below 4 and three (10.3%) remained at or above 7. Unchanged totals did not imply unchanged membership: across the cut-off of 4, one platform moved up and one moved down under the Tree Safety scenario, compared with three in each direction under the export-priority scenario. No platform crossed the cut-off of 7 in either scenario.

The demonstration group had a median overall score of 5.40 (first–third quartiles: 3.80–7.08), compared with 3.82 (2.56–4.40) in the documentation-based group. The two-sided Mann–Whitney comparison yielded U = 138.0, p = 0.084, and r_rb = 0.394. Scores tended to be higher in the demonstration group, although the comparison did not reach the conventional 0.05 significance threshold. This result does not establish equivalence or exclude evidence availability bias.

4. Discussion

4.1. Contribution in Relation to Existing Evaluations

The principal contrast is between relatively widespread mapping and less consistent functions for preserving, interpreting, and acting on inventory data. A review of urban resource tools distinguished basic information provision from forecasting, scenario analysis, and broader management support [16]. Municipal inventory studies addressed updating and operational use rather than software feature maturity [4,5]. The present results extend that context by assessing the supporting software functions directly while not establishing that software limitations caused the organizational patterns described in those studies.

Fry et al. reviewed digital tool applications in urban forestry research and found that snapshots of existing conditions were more common than historical or future-oriented analyses [8]. Our low scores for predictive maintenance, repeated-history functions, and reassessment alerts are conceptually consistent with that imbalance. However, their units were research publications and tool uses, whereas ours were software platforms and sub-criteria. Their percentages should not be treated as benchmarks against which the present percentages can be statistically compared. The new evidence here is the comparison of management-relevant functions within a common 29-platform assessment, not discovery of a previously unknown distinction between mapping and management.

Inventory method reviews and parameter selection studies evaluate which observations can be collected and which information is useful [2,15,17]. General software selection research addresses the translation of requirements into evaluation criteria [12]. The present synthesis connects those perspectives at the platform level through the six domains in Table 1. It is an exploratory assessment structure and a proposed procurement aid, rather than a validated universal maturity standard or a new multi-criteria algorithm.

4.2. Fieldwork, Usability, and Public Access

Mapping reached the usable/strong cut-off in 62.1% of platforms, compared with 44.8% for mobile support and 41.4% for offline operation. This difference supports assessing field workflows separately from spatial display. Image-assisted and street-view approaches can complement field inventory methods [6,7] but do not remove the need to consider the accuracy of the attributes collected or the ability to update records. For procurement, a relevant test is whether staff can create, validate, synchronize, and later retrieve records under the conditions of the intended survey.

The public-access criterion showed a concentrated pattern: seven maximum scores and 22 zero scores. This can reflect specialist platform scope and the broad way the combined viewing/reporting criterion was applied, rather than a smooth gradient of citizen engagement capability. The professional and municipal focus of the purposive sample may also underrepresent tools developed primarily for participation. The data cannot distinguish these explanations or establish a selection effect. Consequently, 24.1% is a description of this sample, not an estimate of public access across all urban tree applications.

Participation additionally requires data quality procedures. Citizen science inventory research highlights the role of protocols, training, and validation [18], and street-level imagery can support some inventory variables more reliably than others [19]. Public viewing, public reporting, and unrestricted editing should therefore be treated as different workflows. Role-based access and review procedures are practical requirements proposed here, not capabilities demonstrated by every high-scoring public-access platform.

4.3. Data Portability and Municipal Data Governance

Online storage and export were not interchangeable strengths: 48.3% reached the usable/strong storage cut-off, compared with 20.7% for export/migration. Long-term monitoring research emphasizes continuity of records across management and governance contexts [20]. A separate study showed how legacy tree records can require geocoding before integration with spatial data [21]. Neither study directly estimates vendor lock-in, but both illustrate why preserving usable records matters beyond the initial survey. Our proposed procurement response is to verify ownership, field-level export, attachment transfer, format documentation, backup, and exit arrangements before committing to a platform.

The hosting/data protection criterion also needs a narrower interpretation than legal compliance. Only eight platforms reached the ≥7 cut-off for the arrangements documented in this assessment. The remaining scores do not establish GDPR violations or insecure servers. Missing or incomplete evidence, an aggregate criterion combining different issues, and the absence of a legal or cybersecurity audit limit that conclusion.

Tree attributes alone need not identify a person, but accounts, staff assignments, citizen reports, photographs, and logs may involve personal data. For such processing, municipalities must identify controller/processor responsibilities and lawful purposes, obtain appropriate processor guarantees and agreements, address subprocessors and international transfers, and establish security, retention, deletion, and breach procedures (GDPR Articles 4–6, 28, 32–34, and Chapter V). Non-compliance can lead to corrective orders or processing restrictions under Article 58; fines for public bodies depend on Member State rules under Article 83(7). Technical export of an inventory is distinct from the personal right to data portability in Article 20, which does not apply to processing necessary for public interest tasks or official authority [22]. These obligations support checking documented governance arrangements, not treating a hosting location or a low software score as a compliance verdict.

4.4. Environmental Interpretation and Tree-Safety Workflows

Environmental tools reached the usable/strong cut-off in 31.0% of platforms, compared with 62.1% for mapping. This distinction concerns integration of analysis, not whether tree data can have environmental value. Knoke et al. discuss converting inventory observations into information for ecosystem service and biodiversity planning [10]. In an Oslo case study, adding geospatial information increased the share of existing tree records suitable for i-Tree Eco analysis from 19% to 54% [23]. That study evaluated data suitability rather than commercial platforms. It nevertheless supports specifying required fields, completeness checks, and spatial linkage when assessing an environmental module.

Reviews of urban remote sensing describe substantial data-processing and validation requirements [24]. The present LiDAR/3D score should therefore not be interpreted as a test of a specific point-cloud algorithm. Attaching a 3D file or showing an external viewer differs from importing georeferenced data, linking them to individual trees, and carrying out measurements or comparisons. Formats such as LAS/LAZ for point clouds and OBJ for models are examples of compatibility questions for future procurement testing; they were not systematically benchmarked in the present study. The functional distinctions clarify the criterion without retrospectively claiming that every analytical task or format was tested.

Tree-safety documentation showed a similar difference between storing information and supporting follow-up. Instrument uploads reached the usable/strong cut-off in 37.9% of platforms, but dynamic history did so in 24.1% and reassessment alerts in 13.8%. Research on visual and instrumental tree assessment supports maintaining interpretable diagnostic evidence [13,14]. Longitudinal inventory research also links repeated tree measurements and management records with the study of growth and tree loss [25]. These sources support the relevance of a connected record but do not independently validate the software scores or show that an attachment alone can support a risk decision.

4.5. Weighting, Evidence, and Procurement Implications

The two alternative priority scenarios changed the aggregate mean by approximately −0.10 and +0.03 points. The counts at the ≥7 cut-off were unchanged, while a single platform crossed ≥5 under the Tree Safety scenario. This supports limited stability of the composite summaries under the two specified scenarios. It does not establish invariance under every weighting scheme, validate individual ratings, or prove that the categories are equally important. The category means themselves remain unchanged by these scenarios because only the overall aggregation weights change.

The demonstration group included platforms accessed through different engagement routes, ranging from routine pre-sales presentations to individually arranged research or professional exchanges. Higher scores in this group may therefore reflect differences in platform scope and provider engagement practices, alongside differences in the evidence available to the assessors. Relative cost categories should not be interpreted as measures of provider size, revenue, or development capacity. Because these organizational characteristics were not measured, the comparison cannot determine their contribution to the observed score differences or separate it from the contribution of evidence availability. Trial/demo availability also contributes to the licensing score, although this criterion is distinct from whether a live demonstration occurred during the study. The comparison therefore remains an exploratory association using the existing composite score, rather than an isolated estimate of the causal effect of demonstration-based assessment.

Table 9 separates our proposed selection requirements from the empirical score tables. Its Essential, High, and Desirable labels are an author synthesis, not priorities elicited from a representative sample of municipalities or an externally validated standard. A municipality should adapt them to inventory purpose, staffing, legal obligations, existing systems, and risk responsibilities. Software fees alone do not capture storage, modules, training, migration, or exit costs; these aspects should be addressed in the specification and quotation [11,12].

Table 9. Proposed requirements checklist for software selection. The checklist synthesizes the assessed functional domains and the authors’ interpretation of their operational relevance. It is not a scored comparison of products, a validated procurement standard, or a set of priorities elicited from municipalities. Essential indicates a proposed prerequisite for basic continuity and control of records; High indicates a proposed operational enhancement; Desirable indicates an additional analytical or integration capability. Local legal or operational requirements may override these labels. GIS: geographic information system; VTA: visual tree assessment; API: application programming interface.

Requirement	Why It Matters	Evidence to Request	Proposed Priority
Data ownership and export rights	Protects the field data investment and reduces dependence on one provider.	Written ownership terms; export of records and attachments in documented, usable formats.	Essential
GIS-based tree mapping	Provides spatially explicit tree records.	Map-based records with coordinates, layers, and editable geometries where required.	Essential
Customizable data fields	Aligns the database with local protocols and future requirements.	Editable fields/forms demonstrated before data collection.	Essential
Mobile field collection	Reduces transcription and supports direct record creation.	Functional phone/tablet workflow for adding and editing records.	Essential
Offline fieldwork	Maintains field operations where connectivity is weak.	Offline capture with tested synchronization and conflict handling.	High
Storage, backup, and recovery	Protects photographs, records, and longitudinal data.	Documented limits, backup frequency, recovery process, and retention period.	Essential
User roles and permissions	Controls access across staff, contractors, and supervisors.	Role-based viewing/editing and an auditable approval workflow.	Essential
Tasks and work orders	Connects inventory records with maintenance actions.	Assignment, status, due dates, completion records, and history.	High
VTA-compatible fields	Supports repeatable tree-safety documentation.	Structured fields/templates for observations, risk ratings, and recommendations.	High
Diagnostic attachments	Links photographs, reports, and instrument outputs to the tree record.	Per-tree attachments with documented capacity and export; separately verify any claim of native analysis.	High
Reassessment scheduling	Supports periodic review and follow-up.	Automatic reminders and scheduled next inspection dates.	High
Customizable reporting	Supports communication with authorities, clients, and stakeholders.	Editable templates, selected fields, organization identity, and exportable reports.	High
Environmental tools	Supports strategic ecosystem service or biodiversity reporting.	Defined inputs, calculation method, treatment of uncertainty, and export.	Desirable
API/interoperability	Enables connection with municipal and GIS systems.	Documented API or tested import/export workflow.	Desirable
Transparent total cost	Avoids underestimating modules, storage, training, support, and exit costs.	Written quotation covering licenses, users, modules, hosting, migration, and retention.	Essential

The common implication is to define the intended post-survey workflow before collecting records at scale: who updates them, how interventions are linked, when reassessment occurs, and how the organization retains usable data after a service change. Earlier municipal studies show that inventories do not automatically become strategic management tools [4,5]. The present platform-level findings add functional questions for procurement, but do not measure the financial loss from selecting an unsuitable system or the management benefit of adopting the checklist.

5. Limitations

The sample was purposive and identified through heterogeneous, non-systematic sources. Products that were less visible or lacked assessable documentation may be underrepresented. The absence of a complete historical query and screening log limits reproducibility of selection. Anonymity further prevents readers from checking product identities and independently reproducing platform-level scores. Aggregate reporting and disclosure of the rubric improve transparency but do not remove this restriction.

Final scoring was performed by one researcher after discussion with a second researcher. Independent blinded ratings were not collected, so neither inter-rater agreement nor the magnitude of individual evaluator bias can be quantified. The anchors structure judgment but do not establish objectivity. Intermediate scores were not calibrated as equal intervals, and means, SE values, and composite scores depend on the chosen numerical coding. The weight sensitivity analysis addresses two aggregation choices only; it does not validate the scale, assess uncertainty in individual ratings, or replace independent replication.

Evidence differed between demonstrated and undemonstrated platforms, and documentation may omit functions that are actually available. The completed numerical matrix does not provide a separate uncertainty classification for every observation. Low assigned scores should therefore be distinguished from a verified absence of all corresponding capabilities. The exploratory group comparison cannot rule out this evidence effect or separate it from differences in platform scope and provider engagement practices. No independent field performance, cybersecurity, legal compliance, or format-by-format 3D benchmark was conducted.

Each platform was assessed once at a different time during March–December 2025. There was no synchronized final re-review, so capabilities may have changed within or after that period. Pricing classes were broad and relative, and incomplete pricing information limited comparability. Future validation should use independently applied ratings, archived evidence, task-based testing, and repeated assessments. Additional municipality-specific weights and clearer separation of capability from evidence certainty would further test the proposed framework.

6. Conclusions

Within this purposive sample of 29 platforms, mapping and core data entry functions were more consistently implemented than strong export/migration, ecological analysis, longitudinal safety records, and reassessment alerts. These findings answer the study’s first two questions at the level of documented platform functionality, without establishing the prevalence or quality of these functions throughout the market.

For the third question, the assessment provides a structured way to define and inspect software requirements before large-scale field collection. The contribution is the application-specific combination of operational, data management, environmental, and safety domains, together with transparent scoring and an explicitly provisional checklist. The two weighting scenarios showed small aggregate changes, whereas the evidence-source comparison left possible bias unresolved.

The framework is a starting point for municipal and professional evaluation, not a universally validated standard. Its usefulness depends on matching requirements to local workflows and recognizing the limits of judgment-based scores. Independent rating validation, realistic field tasks, and periodic reassessment remain necessary to establish stronger reliability and broader applicability.