Figure 1.
Conceptual models of the illegal ivory trade and seizures data.
(A) In each county, in each year an unknown proportion (seizure rate) of illegal ivory transactions (see Figure 1B for examples) is seized. Of these seizures an unknown proportion (reporting rate) are reported to ETIS. We identify potential predictors that discriminate different countries ability to make and report seizures so that we can obtain relative estimates of seizure and reporting rates and the numbers of illegal ivory transactions. Figure S1 shows the predictors of seizure and reporting rate identified by the modelling exercise described in this paper. (B) Flow of ivory from source to point of sale representing different types of illegal ivory transactions. Circles represent individual countries along the trade chain. Raw ivory obtained from illegally killed elephants (or stolen from stockpiles of ivory) is put together into a shipment – this ivory may be sourced from several countries. Shipments might pass through a number of countries before arriving at ivory processing plants. Once the ivory is processed into worked ivory it is put into a shipment (again this ivory may come from several ivory processing plants in different countries and from stockpiles) and could pass through several countries before arriving at wholesale or retail places. Once there, it will likely be sold to individuals, locals or tourists, and could pass through several more countries before reaching its final destination. Seizures can occur all along this trade chain and seized ivory should pass to the ivory stockpiles which also includes ivory from natural elephant mortality in range States. There is evidence that ivory sometimes re-enters the trade chain from unsecured ivory stockpiles [45]. The whole, or parts, of the trade chain for any single piece of ivory could occur in one country, one region or across the globe.
Table 1.
Candidate variables for predictors of seizure and reporting rates.
Figure 2.
Bias adjusted and smoothed expected number of small raw ivory transactions for USA and Cameroon.
(A) Relative seizure (orange) and reporting (blue) rates for Cameroon. Dark shading indicates most likely (median) values and 80% credible intervals are shown. The reporting rate is generally low because reports of illegal ivory seizures are made by NGOs rather than CITES Management Authority (CMA) except in 2004 when the CMA reported to ETIS. Credible intervals for reporting rate are much smaller than for seizure rate. (B) Relative seizure (orange) and reporting (blue) rates for USA. Shading as for A. Reporting rate is high because the USA provides automated and routine reports of illegal ivory seizures to ETIS. (C) Number of reported small raw ivory seizures for USA (purple) and Cameroon (green). (D) Bias-adjusted and smoothed relative numbers of small raw ivory transactions for USA (purple) and Cameroon (green). Dark shading indicates most likely (median) values and 80% credible intervals are shown.
Figure 3.
Transactions Index and Weights Index aggregated over ivory classes and countries with 90% credible intervals.
(A) Transactions Index (TI) and (B) Weights Index (WI). Dot represents the mean and the line the 90% credible interval for each year. The TI and WI are standardised by setting the 1998 value to 100 to constitute a baseline for comparisons with other years. This year was chosen because it was the first full year after ETIS (and MIKE) were mandated and African Elephant populations in three countries moved from CITES Appendix I to Appendix II enabling a tightly-controlled one-off sale of ivory from these countries to Japan to transpire in 1999.
Figure 4.
Transactions Index for each ivory class with 90% credible intervals.
Dot represents the mean, the line the 90% credible interval for each year. For each ivory class the TI is standardised by setting the 1998 value to 100 to constitute a baseline for comparisons with other years. Note the different y-axes due to the relative increase in each ivory class since 1998.
Figure 5.
Mean contribution of each ivory class to the Transactions Index and the Weights Index.
(A) Transactions Index (TI) and (B) Weights Index (WI) are standardised by setting the 1998 total to 100 to constitute a baseline for comparisons with other years.
Figure 6.
The mean Transactions Index for each region within each ivory class.
For each ivory class the Transactions Index (TI) is standardised by setting the 1998 total to 100 to constitute a baseline for comparisons with other years. Mean TI represents the relative number of seizures a country might be expected to make and report to ETIS if seizure and reporting rates were the same over time and across all countries. The countries within each region are listed in Table S2.
Table 2.
Number of seizures in each ivory class.