How Many Real Islands Have Existed along a Mid-Danube Section during the Past 250 Years?

Abstract

The consequences of river regulation work and other human interventions from the past can still be followed and cause various effects on the Danube river, especially on the formation and state of the islands. We aimed to compile an inventory of the islands along the Hungarian Danube stretch between Vének (1797 rkm) and Budapest (1647 rkm) during the past 250 years, focusing on their landscape history. Real islands were defined as gravel bars that are permanently covered with pioneer woody vegetation, constantly surrounded by water (side-branch is not closed), and their surface remains unflooded at an average water level. We identified 123 real islands that existed on this 150 km long river stretch in the last centuries. These were real islands for a shorter or longer period between the 18th and the 21st century. A total of 66 of them existed at the same time about 200 years ago, i.e., before the large-scale water regulation works, while today this number is only 18, meaning that 73% of current potential Danube islands are at a stage of side-branch succession. Before the river regulation works, the natural successional changes of a real island occurred over hundreds of years, but today they happen rapidly. The formation of new islands became very limited compared to the past due to the lack of sediments and the altered river dynamics. In order to conserve this unique ecological corridor and green infrastructure element in the long run, and restore its damaged floodplain habitats, no more hard-tech interventions should be allowed in the fluvial system. These aspects need to be taken into consideration in decision-making processes with an integrated approach.

1. Introduction

The Danube is the most international river in the world [1], and Europe’s second-largest. Today, many stretches of the whole river are directly influenced by dams. The Hungarian stretch has remained free-flowing, which is of unique value, but during the last centuries, many other artificial interventions have harmed the river, causing significant effects on its islands and their habitats [2]. In Hungary less than 3% of the former fluvial forests have been preserved [3]; this shows the scale of harmful interventions and land-use changes [4]. The floodplain habitats still represent high ecological value for both biodiversity and humans.

The formation of gravel bars and islands as well as the meandering characteristics of the river depend on the water-level gradient and runoff [5]. In order to explore the time horizon of geomorphological processes, literature regarding terrace-formation during the age of the Danube riverbed and its geomorphology should be reviewed. Late Pleistocene and Holocene climatic changes led to the development of the Danube floodplain and the formation of the islands. During the Atlantic–Boreal transition and the first third of the Atlantic, an upper floodplain had been formed about 1–2 m above the lower floodplain that composes part of the islands. The more intensive riverbed incision led to the formation of the lower floodplain during the second half of the Subboreal and after it (late Iron Age) [6,7]. As a consequence of the incision, riverbeds that surround the islands become well located and several small gravel bars and islands merged.

The development of the bars was caused by the locally decreased flow energy, which leads to the deposition of coarse bed-load sediment and fine suspended sediment also trapped behind them [5]. The deposited sediment splits the flow and causes lateral bank erosion [8]. The base of a gravel bar is created by floods, while it is further shaped by lower waters [9,10]. As part of studying river islands, it is logical to define the island as having a fluvial morphological character. According to its general definition, an island is an area of land smaller than a continent and entirely surrounded by water [11]. From a geomorphological perspective, river islands are viewed as the passive result of the physical process regime of fluvial systems, covered with vegetation [12]. Pioneer floodplain vegetation stabilizes the gravel bars and thus, plays a crucial role in the formation of the islands [13]. Bio- and geo-morphological processes that shape the landforms of the fluvial systems (e.g., gravel bar, island) are constantly in motion [10,14]. The existence of gravel bars and their dynamic surface changes are sensitive indicators of human influence in fluvial morphology [15]. The value of floodplain landscapes is increasingly determined by their water supply [16]. Connectivity enhancement has also become a conservation priority, both longitudinal and lateral [17].

The objectives of this study are to gain knowledge about the formation, change, and disappearance of the islands along the Hungarian Danube stretch between Vének and Budapest. Integrating data from historical maps to see the various transformations and shifts of the riverbanks (e.g., [18]) is of high importance in order to provide a perspective on the required land management and necessary interventions to maintain them sustainably. This is why we aimed to explore how many islands existed during historical times, follow their natural changes, and also those changes which are determined by artificial interventions, and to see how many islands are formed or disappeared compared to the past and which ones could survive as real islands till nowadays. We have not found any other studies that observe the long-term effects of human interventions on the river islands via the analysis of historical maps.

The cumulative effects and the interactions of the natural processes and artificial interventions in line with the extreme climate conditions, rapid land-use change, hydropower plants, eutrophication due to excess fertilizer use, and the expansion of invasive alien species have caused extremely negative effects on the islands never been experienced before [19,20].

2. Materials and Methods

The Danube enters the Carpathian basin at Devin (Slovakia). The velocity of the stream slows down and creates a special inner delta in the Szigetköz (Hungary)—Žitný ostrov (Slovakia) region. After the mouth of the Mosoni Danube branch, the riverbed is more stable; therefore, we started our studies from this site, i.e., Vének (1797 river kilometre = rkm) (Figure 1), and moved along the main Danube stretch towards Budapest (1647 rkm). We divided this 150 rkm long stretch into three sections as it is also used in mapping sources.

• Vének—Komárom (1797–1766 rkm)—31 rkm
• Komárom—Esztergom (1766–1721 rkm)—45 rkm
• Esztergom—Budapest (1721–1647 rkm)—74 rkm

During the landscape historical investigations, we studied the Hungarian geographical researches and references [21,22,23,24] to collect the list of potential mappings about the Danube from the island’s point of view. The first “map” (the “Tabula Peutingeriana”) where the Danube is presented by a line dates back to the Roman Age [25]. During the 14th century, some portolan maps also appeared, but the first geographically valid map for the Danube originated at the end of the 17th century. This is the Danubius Pannonico Mysicus, created by Luigi Ferdinando Marsigli [26]. Other important geographical references were the county maps by Lipszky (1799 and 1806), the navigation map of Bänhöltzel (1771) and Pasetti (1857), the first technically well-developed Danube survey, the military surveys, the river regulation plans (including their maps), the maps for water sports, the topographical maps, and aerial and satellite photos.

During the searching work of the available maps and surveys, we visited the following institutions: Map Archive of the Military History (Budapest); National Archives of Hungary; Danube Museum (Esztergom); Special Library of Water and Environmental Protection (Budapest); county and town libraries, local historical collections and second-hand bookshops. The online hungaricana.hu (Hungarian Cultural Heritage Portal), mapire.eu (Historical Maps Online) and fentrol.hu (Online Aerial Photo Archive) webpages contributed greatly to the remote access of a huge amount of sources as well.

As part of the landscape historical research, we visually analyzed 29 mappings from the last 300 years. We present the analyzed results of 10 mappings in the case of each section, which are presented in 20 to 30 interval years. Not every mapping covered the whole 150 rkm-long section, and some map sections were lost during the past centuries, or in some cases we considered the analysis of another mapping are from a certain period to be more suitable for our aim (

Table 1. Analyzed maps for each Danube section and their reference numbers or sources, and scale.

Time of Mapping	Name of Mapping	Reference Number or Source in the Map Archive of the Military History	Scale	Vének-Komárom	Komárom-Esztergom	Esztergom- Budapest
169x	Marsigli Danubius P. Mysicus	own		x	x	x
1771	Navigation map of Bänhöltzel	S12-Div.X.-No.92	1:12,240	---	XX	XX 3
1782–1785	1st military survey	DVD	1:28,800	XX	x	x
1799	Manuscript of Lipszky: Mappa Generalis	B IX a 518/b	1:230,400	x	x	x
1806	Lipszky: Mappa Generalis	B IX a 518	1:480,000	x	x	x
18xx	Bécs-Orsova-Ruse	B IX b 128/1	1:72,000	x	x	x
1809	Mapping of Xivkovich	B IX b 401	1:14,400	XX	---	---
1830–1840	Danube survey	S80-Duna-No.126	1:3600	XX	XX	x 1
1834	Map named after Rauchmüller	B IX b 134	1:36,000	x	x 1	x 1
183x	Danube map	B IX b 136	1:14,400	x	XX	x
1836–1840	Topographical map of the Danube	B IX b 122/2	1:36,000	x	x	XX
1842	Hummitzsch (Vienna–Budapest)	B IX b 122		x	x	x
1857	Navigation map of Pasetti	B IX b 138	1:28,800	XX	XX	XX
1869–1887	3rd military survey	DVD	1:25,000	XX	XX	XX
1910	Danube map	B IX b 155a	1:5000	x	XX	x
1911	Danube map	B IX b 157/1	1:25,000	XX	x	XX
1922–1935	Water sports map (‘with angels’)	own	1:25,000	XX	XX	XX
1930	Rowing map of Kemény Andor	own	1:25,000	---	---	x
193x	Danube map	B XV b 181	1:10,000	---	---	x 4
195x	Danube map	B XV b 142	1:8500	---	---	x 4
1958	Water sports map	own	1:40,000	---	---	XX
1965	Topographical map	B XV b 343	1:10,000	XX	XX	XX
1975	Water sports map	own	1:40,000	---	---	x
1980–1990	Topographical map	own	1:10,000	XX 2	XX 2	XX
1985	Water sports map	own	1:20,000	---	---	x
1992	Yachting map	own	1:50,000	x	x	x
2005	Orthophoto series of Hungary	Ministry of Agriculture	1:10,000	x 2	x 2	x
2015	Orthophoto series of Hungary	Ministry of Agriculture	1:10,000	x 2	x 2	x
2018–2019	Maxar satellite images	ESRI Basemap	1:10,000	XX	XX	XX

Notes: XX = presented in the summarizing Table 2, Table 3 and Table 4; x = analyzed only; ¹ = map sections lost; ² = only the Hungarian side is visible; ³ = until Nagymaros; ⁴ = from the mouth of the river Ipoly. Source: authors’ elaboration.

).

During the analysis of each map, we summarized the number of islands and gravel bars by location in each period. Besides these usual natural river formations, we differentiated those transformation types that evolved (at least partly) due to artificial interventions in order to describe their actual state. We did not find any example for such an alignment in the literature. The river functional unit approach of Amoros et al. [27] refers to the silting up of the branches, but does not focus on islands, and leaves the artificial factor unattended, which we considered to be important. This is why we created a category system from the perspective of the islands and the affected artificial interventions to describe the actual state of the islands on each map. We summarized the results by the river sections mentioned above (see Table 2, Table 3 and Table 4).

• Gravel bar (GB in Table 2, Table 3 and Table 4)—deposited sediment in the riverbed, without vegetation. Its presence on a map depended on the purpose of the mapping and also on the water level; therefore, uncertainty appeared. Apart from this weakness, these maps are still useful to detect most of the gravel bars because many times the presence proves that the gravel bar, which is the basis of a latter island, was formed more than a hundred years ago.
• Real island (RI in Table 2, Table 3 and Table 4)—is born when the pioneer woody vegetation is established on a gravel bar, surrounded by a water branch. Its size and shape will be formed by local hydrodynamic conditions. Its surface remains unflooded in case of average water level.
• Island with closed side-branch (CB in Table 2, Table 3 and Table 4)—the island’s side-branch is closed by a rock-fill dam, the hydrodynamics of the side-branch changes, the flow of the water decreases and induces siltation, but this process is still in the early phase.
• Island with drying out side-branch (DO in Table 2, Table 3 and Table 4)—the silting-up is in the intermediate phase, which means that the side-branch temporarily dries out when the water level is very low. It has clear signs on the maps.
• One-time island (OTI in Table 2, Table 3 and Table 4)—the silting-up is in an advanced phase and the island practically becomes part of the riverside. (NB: those islands whose surface does not exist anymore (were totally dredged up or disappeared due to river dynamics) do not belong here.)
• Island with re-opened side-branch (RO in Table 2, Table 3 and Table 4)—an additional category for those islands where the side-branch had been physically closed earlier, but has been recently re-opened.

A difficulty of riverine mapping is that the objects (unflooded surface of gravel bars, drying out side-branches, etc.) seen on the spot and shown on the map significantly depend on the actual river water level. Moreover, the elaboration of each mapping, the measurements used, and thus, the extent of showing the reality depends on the exact aim of the mapping (military, navigation, topography, tourism, sport, etc.) as well as the technique used, and varies across the centuries. We tried to minimize the potential risks emerging from these factors by performing the following:

• We observed more than one map from each period in order to filter out the bias caused by water-level changes.
• We focused on creating obvious categories (differences between gravel bar and island, stages of silting up in side-branches, which still contains some subjectivity).
• We considered the time of creation or data capture, as the period throughout the year could massively influence the analysis.
• Navigation, hydrology and water regulation maps show the most precise picture from the island morphological aspect; therefore, they received greater emphasis.
• By using the large scale of the chosen mappings we ensured detailed imagery.

In order to present the river dynamics that create or ruin islands, we included those small or very new islands in Table 2, Table 3 and Table 4 that were not named on any map or in the literature. We gave names to them based on their characteristics and indicated these cases with ON = own name in the tables. If two or more islands merged, we indicated this by merging the cells of the tables. Geographical names throughout the paper are used in their original form (on the referred map), adding their current official name in brackets if it is different.

The georeferencing of some maps, satellite/aerial photos and the GIS measures of the islands (size, shape) were prepared by the use of ESRI ArcGIS 10.4.1 for Desktop software. Figures were created by Microsoft Excel software.

3. Results and Discussion

3.1. Vének–Komárom (1797–1766 rkm) River Section

As already mentioned, the velocity of the Danube slows down after entering the Carpathian basin, leading to the formation of a special inner delta in the Szigetköz (Hungary)—Žitný ostrov (Slovakia) region. The riverbed is considered more stable after the mouth of the Mosoni Danube branch (this is why we started our studies from this site, i.e., Vének). However, our landscape history studies show that this stability is surprisingly “new”. Before the investigated centuries the main branch had been moving more intensively, the evidence of the ancient branches can be followed on the maps and also was found in medieval codex notes. Water regulation interventions of the 19th–20th century caused narrower floodplains and regulated main branches, leading to silting up in the side-branches of the islands [28]. Those one-time islands between Vének and Gönyű may have been the last remnants of the ancient, braided riverbed; the river changed onto a meandering–silting up character from here. Today this riverbed is not able to meander anymore due to the longitudinal training structures and river bank defense with ripraps against lateral erosion, and thus, the memento of past hydrodynamic processes are the old islands with closed or drying out side-branch between Gönyű and Komárom.

The greatest number (20 to 21) of islands on the 31 km-long Vének-Komárom Danube section (Table 2) was shown in the mid-19th century (Danube survey, Navigation map of Pasetti). None of them remained real islands to date; the side-branch of one of them (Nagy Lél = Veľkỳ Lel island) was partly re-opened recently as part of a rehabilitation project. Currently, two further islands can be now evaluated as real islands (i.e., thoroughly surrounded by the river), both of which are relatively young (50–80 years old) and small (<5 ha).

Table 2. Changes of the islands on the Vének—Komárom Danube section based on mappings. Grey background highlights the real islands. Blank place means that no island or gravel bar existed at that time.

		Year of Mapping	1782–1785	1809	1830–1840	1857	1869–1887	1911	1934 1922–1923	1965	1980–1990	2018–2019
Settlement	river km	Number on Figure 1. Name of the island	1st military survey 1:28,800	Xiv-kovich 1:14,400	Danube survey 1:3600	Pasetti 1:28,800	3rd military survey 1:25,000	Danube map 1:25,000	Water-sports map (Angel) 1:25,000	Topogr. map 1:10,000	Topogr. map 1:10,000	Maxar satellite ESRI
Vének	1797	1. side-branch i. ON									CB	CB
		2. side-branch i. ON									CB
		3. pre Kolera ON							GB	CB	CB
	1795	4. Kolera			GB	RI	RI	CB	CB	CB
	1794	5. Torda	RI	RI	RI	RI	RI	CB	CB	CB	CB	CB
Klížska Nemá	1795–1793	6. Merítő			GB	RI	OTI	OTI	OTI	OTI	OTI *	OTI
		7. pre Old ON			GB	RI
		8. gravels			RI	GB
		9. Old	RI		RI	RI
		x Changing ON	RI	RI
	1794	10. Kitty ON					GB	GB	RI
	1793	11. Ásványtő ON					RI	RI	GB			CB
	1792	12. Vándor ON							GB	RI	RI	CB
	1792	x Appeared ON									DO
	1789	13. Némai	RI	RI	OTI	OTI	OTI	OTI	OTI	OTI	OTI *	OTI
	1791	x Sarkantyús 1									*	DO
	1790	x Sarkantyús 2									GB	GB
	1790	14. Tómellék Up ON							GB	RI	RI	RI
	1789	15. Tóm. Down ON								RI	RI
Gönyű, Nagyszentjános	1787–1789	x gravel bar								GB	GB	DO
		16. Young pre ON								RI	RI
		17. Small E. pre ON	RI **	RI	RI	RI		RI	RI	RI	RI	DO
		18. Small Erebe		RI	RI	RI	RI
		19. Great E. pre ON	RI **	RI	RI	RI
		20. Great Erebe			RI	RI	RI	RI	RI	RI
		x side-branch bar				GB			GB	CB	CB
		x l. training dam 1								CB	CB
		21. E. side-branch i.						GB	RI	DO	DO
		22. Cat				GB	RI	CB	CB	CB	CB	DO
		x Lower bar							GB
		x l. training dam 2				GB				DO	DO
		x l. training dam 3								DO	DO
		x l. training dam 4									DO
		23. i. near Cat ON									RI
Kkeszi	1786	x Kiskeszi bar ON				GB	GB		GB	GB	GB	GB
		24. Opposite bar ON				GB			GB	RI	RI	DO
Veľkỳ Lel	1780–1786	25. VL. Vesszős	RI	RI	RI	RI	RI	CB	CB	CB	CB *	RO
		26. VL. Kis		RI	RI	RI	RI
		27. Velky Lel		RI	RI	RI	RI	CB	CB	CB	CB *
		28. side-branch Upper			RI	GB					*
		29. side-branch Lower	RI					RI	RI		*
		30. Hosszúkás ON			GB	GB	GB	OTI	OTI	RI	*
		31. post VL ON	RI	RI	RI	GB		GB	RI	RI	RI*
Ács	1783–1784	32. pre Zsidó								RI	RI	RI
		33. Zsidó				GB	GB	GB	RI
	1779	34. Ács bar			RI	RI	DO	OTI	OTI	OTI	OTI	OTI
	1777	35. Concó ON				GB		CB	CB	CB	RI	OTI
Komárom	1776	36. Lídia islands/bar	GB	RI	RI	RI			GB	OTI	OTI *	OTI
					RI	RI
					RI	RI					*
	1775	x narrow bar				GB
	1775–1773	37. SzP. side-branch ON	RI		GB	RI	CB	CB	CB	CB	DO	DO
		38. Szent Pál	RI	RI	RI	RI	CB	CB	CB	CB	CB	DO
	1774	39. SzP. Small ON	RI	RI	RI	RI	CB	CB	CB
	1772	40. Monostori	RI	RI	RI	RI	CB	CB	CB
	1770–1767	41. Erzsébet	RI	RI	RI	RI	RI	CB	CB	CB	CB *	CB
		SUMMARY
	gravel bar (GB)		1	0	5	12	4	4	9	1	3	2
	real island (RI)		14	15	20	21	10	4	7	10	10	2
	closed side-branch (CB)		0	0	0	0	4	11	11	12	11	5
	drying out side-branch (DO)		0	0	0	0	1	0	0	3	6	7
	one-time island (OTI)		0	0	1	1	2	4	4	4	4	5
	re-opened (RO)		0	0	0	0	0	0	0	0	0	1
	Year of mapping:		1782–1785	1809	1830–1840	1857	1872–1884	1911	1934 1922–1923	1965	1980–1990	2018–2019
	Time intervals (in years):		20–30	20–30	20–25	20–30	20–30	10–15	30–40	20–25	30–40

Notes: Island is abbreviated as i. GB—gravel bar, RI—real island, CB—island with closed side-branch, DO—island with drying out side-branch, OTI—one-time island, RO—island with re-opened side-branch. * Data from the missing section of the topographical map were replaced by maps from a sports navigation book originating from the same period; ** The +2 islands are the Small and Great Erebe islands that are missing from the map due to press mistake (other maps from this period show them.) Source: authors’ elaboration.

Table 2. Changes of the islands on the Vének—Komárom Danube section based on mappings. Grey background highlights the real islands. Blank place means that no island or gravel bar existed at that time.

		Year of Mapping	1782–1785	1809	1830–1840	1857	1869–1887	1911	1934 1922–1923	1965	1980–1990	2018–2019
Settlement	river km	Number on Figure 1. Name of the island	1st military survey 1:28,800	Xiv-kovich 1:14,400	Danube survey 1:3600	Pasetti 1:28,800	3rd military survey 1:25,000	Danube map 1:25,000	Water-sports map (Angel) 1:25,000	Topogr. map 1:10,000	Topogr. map 1:10,000	Maxar satellite ESRI
Vének	1797	1. side-branch i. ON									CB	CB
		2. side-branch i. ON									CB
		3. pre Kolera ON							GB	CB	CB
	1795	4. Kolera			GB	RI	RI	CB	CB	CB
	1794	5. Torda	RI	RI	RI	RI	RI	CB	CB	CB	CB	CB
Klížska Nemá	1795–1793	6. Merítő			GB	RI	OTI	OTI	OTI	OTI	OTI *	OTI
		7. pre Old ON			GB	RI
		8. gravels			RI	GB
		9. Old	RI		RI	RI
		x Changing ON	RI	RI
	1794	10. Kitty ON					GB	GB	RI
	1793	11. Ásványtő ON					RI	RI	GB			CB
	1792	12. Vándor ON							GB	RI	RI	CB
	1792	x Appeared ON									DO
	1789	13. Némai	RI	RI	OTI	OTI	OTI	OTI	OTI	OTI	OTI *	OTI
	1791	x Sarkantyús 1									*	DO
	1790	x Sarkantyús 2									GB	GB
	1790	14. Tómellék Up ON							GB	RI	RI	RI
	1789	15. Tóm. Down ON								RI	RI
Gönyű, Nagyszentjános	1787–1789	x gravel bar								GB	GB	DO
		16. Young pre ON								RI	RI
		17. Small E. pre ON	RI **	RI	RI	RI		RI	RI	RI	RI	DO
		18. Small Erebe		RI	RI	RI	RI
		19. Great E. pre ON	RI **	RI	RI	RI
		20. Great Erebe			RI	RI	RI	RI	RI	RI
		x side-branch bar				GB			GB	CB	CB
		x l. training dam 1								CB	CB
		21. E. side-branch i.						GB	RI	DO	DO
		22. Cat				GB	RI	CB	CB	CB	CB	DO
		x Lower bar							GB
		x l. training dam 2				GB				DO	DO
		x l. training dam 3								DO	DO
		x l. training dam 4									DO
		23. i. near Cat ON									RI
Kkeszi	1786	x Kiskeszi bar ON				GB	GB		GB	GB	GB	GB
		24. Opposite bar ON				GB			GB	RI	RI	DO
Veľkỳ Lel	1780–1786	25. VL. Vesszős	RI	RI	RI	RI	RI	CB	CB	CB	CB *	RO
		26. VL. Kis		RI	RI	RI	RI
		27. Velky Lel		RI	RI	RI	RI	CB	CB	CB	CB *
		28. side-branch Upper			RI	GB					*
		29. side-branch Lower	RI					RI	RI		*
		30. Hosszúkás ON			GB	GB	GB	OTI	OTI	RI	*
		31. post VL ON	RI	RI	RI	GB		GB	RI	RI	RI*
Ács	1783–1784	32. pre Zsidó								RI	RI	RI
		33. Zsidó				GB	GB	GB	RI
	1779	34. Ács bar			RI	RI	DO	OTI	OTI	OTI	OTI	OTI
	1777	35. Concó ON				GB		CB	CB	CB	RI	OTI
Komárom	1776	36. Lídia islands/bar	GB	RI	RI	RI			GB	OTI	OTI *	OTI
					RI	RI
					RI	RI					*
	1775	x narrow bar				GB
	1775–1773	37. SzP. side-branch ON	RI		GB	RI	CB	CB	CB	CB	DO	DO
		38. Szent Pál	RI	RI	RI	RI	CB	CB	CB	CB	CB	DO
	1774	39. SzP. Small ON	RI	RI	RI	RI	CB	CB	CB
	1772	40. Monostori	RI	RI	RI	RI	CB	CB	CB
	1770–1767	41. Erzsébet	RI	RI	RI	RI	RI	CB	CB	CB	CB *	CB
		SUMMARY
	gravel bar (GB)		1	0	5	12	4	4	9	1	3	2
	real island (RI)		14	15	20	21	10	4	7	10	10	2
	closed side-branch (CB)		0	0	0	0	4	11	11	12	11	5
	drying out side-branch (DO)		0	0	0	0	1	0	0	3	6	7
	one-time island (OTI)		0	0	1	1	2	4	4	4	4	5
	re-opened (RO)		0	0	0	0	0	0	0	0	0	1
	Year of mapping:		1782–1785	1809	1830–1840	1857	1872–1884	1911	1934 1922–1923	1965	1980–1990	2018–2019
	Time intervals (in years):		20–30	20–30	20–25	20–30	20–30	10–15	30–40	20–25	30–40

Notes: Island is abbreviated as i. GB—gravel bar, RI—real island, CB—island with closed side-branch, DO—island with drying out side-branch, OTI—one-time island, RO—island with re-opened side-branch. * Data from the missing section of the topographical map were replaced by maps from a sports navigation book originating from the same period; ** The +2 islands are the Small and Great Erebe islands that are missing from the map due to press mistake (other maps from this period show them.) Source: authors’ elaboration.

Some highlights (Table 2)

The first island among the most interesting ones in this section is the Kolera (Cholera) island. Its gravel-bar base emerged during the 1830s. The meander cutoff above this section (between Vének and Nagybajcs) in 1805 played a significant role in its formation. This island is characterized by a classic spindle-shaped core bar. It lost its natural real island character around the late-1890s when a longitudinal training structure was built and later the side-branch was closed in order to regulate low waters. The name Kolera (Cholera) refers to the great cholera epidemic in 1831, when a watch post operated here in order to prevent its spread from the northern bank of the river [29]. The most important group of ship mills (boat mills) operated here, which also emphasizes the power of current flow. The island increased after the closing of the side-branch. Gravel-bar development still continues along its mainstream side shore.

The lower neighbor of the Kolera (Cholera) island is the Torda island. The shape of the Torda island is definitely unusual, which refers to its origin. It had originally been a peninsula of the Vének bank but became progressively detached until the second half of the 18th century. The Torda island had to face the challenges from the currents of both the Mosoni Danube branch and the Old Danube and thus, its size decreased to a quarter of its original area during two centuries. Its side-branch towards the Vének bank was closed to the Kolera island by the previously mentioned longitudinal training structure (the 1890s) and another rockfill dam attached the island to the Vének bank directly. The island and its land use (mainly pasture) appear in numerous old documents [30]. Mining of the island’s material started in the early 1990s to serve the creation of the Győr–Gönyű harbor. Finally, its last natural remnants disappeared in 2018, when the construction of the significant water regulation structure was built to retain water in the Mosoni Danube, as an engineering answer for the riverbed incision caused by the upper hydropower plants.

The first real island along the river flow is named, by us, as the Tómellék Upper island (1790 rkm). It originates from the middle of the 20th century due to the altered flow dynamics. Another proximal island (Tómellék Lower island) appeared directly east of the Upper in the 1950s, but its dredging led to its disappearance until 1990. The Upper island showed the classic spindle shape, characteristic of the bar islands in the main stretch even in the 1990s. It lies behind a new groyne today, and dredging its shape caused a huge loss (about two-third) of its area.

There is evidence for the existence of the Erebe island group from the 13th century [31], and it originally consisted of two islands. The ford at Gönyű was one of the worst from a navigation point of view [32]; therefore, a longitudinal training structure was built in order to detach its side-branch in the 1890s [33]. The current slowed down behind this structure and thus, led to the formation of newer bars and newer islands. Today, the numerous smaller and bigger islands of the Erebe group merge in case of low water level, showing a massive three-times increase.

There is evidence for the existence of the Nagy Lél (Veľkỳ Lel) island group from before the 13th century [34], and it was originally comprised of one great and several small islands around it. The upper end of its side-branch was closed around 1900, leading to its silting-up and thus, the smaller islands merged with the great one, forming practically one big island in the second half of the 20th century. The side-branch silted up but was later partly re-opened after dredging in a recent habitat rehabilitation project.

The gravel bar basis of the Zsidó island (1783–1784 rkm) has been constantly covered by vegetation since the early 20th century. Its theoretical value is that its basis is of natural origin, i.e., any nearby water regulation works only accelerated its island formation process, in line with lowering the water level during average and drought periods caused by the deepening of the riverbed, which still affects the silting-up process in its side-branch and thus, increases the island’s surface. The Zsidó island is a good example of 20th-century island formation. Relevant literature [35] states that the upper end of the islands deteriorates due to the river current, while the lower end constantly develops, causing a downwards movement of the islands. On the contrary, the Zsidó island develops upwards, i.e., against the river current. A possible reason for this might be the deepening of the riverbed and as its consequences, those gravel bars that lie farther from the main current remain unflooded for most of the year due to the lower water level, providing space for the appearance of pioneer vegetation, which stabilizes the surface. The side-branch is narrowing and shallow, which means that the existence of this real island is uncertain in the longer term.

The one-time Lídia island was developed from a left-bank point bar, and was totally removed as part of the river regulation works between 1882–1884 [36]. A new bar started to develop in its place a little later on, but instead of island development, it merged into the bank.

The Koppánymonostori island was once a group of four islands, which were interconnected by training structures after closing their side-branch in the 1880s. They merged until the mid-20th century. The one-time shape of each island can still be recognized. Its side-branch has significantly silted up for now.

Once consisted to be made of two parts, the Elisabeth island at Komárno is a typical streamlined spindle-shaped main-river island. The previous inner side-branch (that silted up) was marked as a ditch even by early 20th century maps. The side-branch was closed with three cross-structure between 1900 and 1902 and converted to a winter harbor. This closing soon became a wide route from the bank into the island.

All the remaining small islands on the Vének–Komárom Danube stretch have existed only temporarily during the past 250 years. They usually evolved as a result of the altered current stretches, or due to training structure, and merged with them and later with the bank.

Key points:

• All studied islands = 41
• Real islands in 2018 = 2 + 1 re-opened; (real islands in 1830s = 20)
• Islands with closed side-branch = 10 (+15 with silted up or drying out branch)
• Islands that merged to neighbor island or into the river bank or their side branch is drying out with or without regulation structures = 25

Number of disappeared or dredged islands = 3

3.2. Komárom–Esztergom (1766–1721 rkm) River Section

The riverbed is stable on this stretch, the river current is balanced now and has been balanced for a long time, as there had been no river regulation interventions before the studied period. The riverbed widens in some places, where the broader river created fords and island groups. The Tát–Ebed–Szentkirály ford was considered as developed and thus, used for river crossing as early as the Roman times [37], giving way to six islands (not at the same time); the Neszmély–Süttő stretch was quite similar, with five islands. The ford at Nyergesújfalu is smaller and its marly–rocky riverbed causes severe difficulties for navigation, hosting one or two islands. An much smaller one was the so-called Szőny-Harcsási ford, blamed for stopping the ice [28,38].

Along this 45 rkm long Danube stretch the greatest number of islands at the same time, namely 21, is shown by the 3rd military mapping from the late-19th century (Table 3). Only one of them, namely the Csitri island meets the requirements of being a real island still today, although its side-branch towards the Körtvélyes island is subject to drying out at a low water level. The other real island is even smaller and much younger. This is the Tiny island (own name), which has been developing for about 30–50 years after the Szőnyi island downstream, directly behind its training structure close to the riverbank, now host to four or five trees, though its development is driven by both natural processes and secondary reasons. All the other side-branches on this Danube stretch have been closed by now—if not directly, then the cross-structure of the adjacent island led to the silting up of the side-branch (e.g., Körtvélyes island, Karvai island).

Table 3. Changes of the islands on the Komárom—Esztergom Danube section based on mappings. Grey background highlights the real islands. Blank indicates that no island or gravel bar existed at that time.

		Year of Mapping	1771	1830–1840	183x	1857	1869–1887	1910	1934 (1922–1923)	1965	1980–1890	2018–2019
Settlement	rkm	Number on Figure 1 Name of the island	Bän-höltzel nav. map 1:12,240	Danube survey 1:3600	Danube map 1:14,400	Pasetti 1:28,800	3rd military survey 1:25,000	map 1:5000	Water sports map (Angel) 1:25,000	Topogr. map 1:10,000	Topogr map 1:10,000	Maxar satellite ESRI
Komárom-Szőny	1764–1762	42. Szőnyi 43. Szőnyi side-branch (ON)	GB		GB	GB	CB	CB	CB	CB	CB	CB
									CB	CB	CB
										CB	CB
										CB	CB
										CB
	1762	44. Tiny ON							GB			RI
		x Izsai bar ON							GB		*
Almásfüzitő	1757	45. pre Prepost ON			RI			RI			CB	CB
	1756	46. Prépost	RI	RI	RI	RI	RI	RI	RI	RI
		x gravel bar						GB	GB
	1755	x Dots ON							GB		*	GB
Neszmély	1751	47. Zsitva mouth	GB		GB			GB	RI		RI *	GB
	1749	48. pre Upper ON		GB	RI	GB	RI	CB	CB	CB	CB	CB
		49. Orphanage					RI	RI	RI	RI
	1748–1746	50. Upper	RI	RI	RI	RI	RI	CB	CB	CB	CB	CB
		x pre Lower bar		GB	GB	GB		GB				CB
		51. Lower	RI	RI	RI	RI	RI	RI	RI	RI	CB
		x pre Radvanyi bar	GB
		x gravel bar	GB
		52. pre Radv. ON		*						RI	RI	CB
		53. Radványi 1	RI	RI	RI	RI	RI	CB	CB	CB	CB
	1745	54. Mocsi 1	RI	RI	RI	RI	RI	CB	CB	CB	CB	CB
Süttő	1743	55. Süttői 1	RI	RI	RI	GB	RI	RI	CB	RI	RI *	CB
	1742	56. Vesszős		GB	RI	GB	RI	RI	CB	OTI	OTI *	OTI
		57. Karvai	RI	RI	RI	RI	RI	RI	RI
	1740	58. Long land ON						RI	RI	OTI	OTI *	OTI
Nyergesújfalu	1734	59. Csenke ON		RI	RI	RI	RI	RI	RI		RI	DO
	1734–1733	60. Újfalusi	RI	RI	RI	RI	RI	RI	RI	RI	RI	CB
		x Ford at Nyerges	GB		GB	GB	GB		GB	GB
Muzsla	1732–1730	x Muzslai bar	GB			GB				DO	OTI *	OTI
		61. Side-br. M. ON		RI		RI
		62. Muzslai	RI	RI	RI	DO	RI	CB	CB
		63. Small Muzsla ON		RI	GB	RI	RI	OTI	GB
Tát	1728–1722	64. Táti	RI	RI	RI	RI	RI	RI	CB	CB	CB	CB
		65. Körtvélyes	RI	RI	RI	RI	RI	RI	RI	RI	RI
	1724	66. Turán	RI		RI	RI	RI	RI	RI	RI	DO
	1723	67. Csitri		RI	RI		RI	RI	RI	RI	RI	RI
	1725–1722	68. Small Poplar ON			RI		RI	RI	GB	CB	CB	CB
		69. Poplar	RI	RI	RI	RI	RI	RI	CB	CB	CB
		70. New Ebedi ON			RI	GB	RI	RI	RI	RI	RI *	OTI
	SUMMARY
	gravel bar (GB)		6	3	5	8	1	3	7	1	0	2
	real island (RI)		13	16	20	14	21	17	12	10	8	2
	closed side-branch (CB)		0	0	0	0	1	6	11	12	13	11
	drying out side-branch (DO)		0	0	0	1	0	0	0	1	0	1
	one-time island (OTI)		0	0	0	0	0	1	0	2	3	4
	re-opened (RO)		0	0	0	0	0	0	0	0	0	0
	Year of mapping:		1771	1830–1840	183x	1857	1869–1887	1910	1934 1922–1923	1965	1980–1990	2018–2019
	Time intervals (in years):		60–70	0–10	20–25	20–30	20–30	10–15	30–40	20–25	30–40

Notes: GB—gravel bar, RI—real island, CB—island with closed side-branch, DO—island with drying out side-branch, OTI—one-time island, RO—island with re-opened side-branch. * Data from the missing section of the topographical map were replaced by maps from a sports navigation book originating from the same period, relevant page of the Danube survey was lost, but its content can be replaced from the Rauchmüller series; ¹ = These names have changed on the latest topographical nap: Radványi = Mocsi; Mocsi = Süttői; the earlier Süttői island belongs to Slovakia on this map and therefore it is not presented. Source: authors’ elaboration.

Table 3. Changes of the islands on the Komárom—Esztergom Danube section based on mappings. Grey background highlights the real islands. Blank indicates that no island or gravel bar existed at that time.

		Year of Mapping	1771	1830–1840	183x	1857	1869–1887	1910	1934 (1922–1923)	1965	1980–1890	2018–2019
Settlement	rkm	Number on Figure 1 Name of the island	Bän-höltzel nav. map 1:12,240	Danube survey 1:3600	Danube map 1:14,400	Pasetti 1:28,800	3rd military survey 1:25,000	map 1:5000	Water sports map (Angel) 1:25,000	Topogr. map 1:10,000	Topogr map 1:10,000	Maxar satellite ESRI
Komárom-Szőny	1764–1762	42. Szőnyi 43. Szőnyi side-branch (ON)	GB		GB	GB	CB	CB	CB	CB	CB	CB
									CB	CB	CB
										CB	CB
										CB	CB
										CB
	1762	44. Tiny ON							GB			RI
		x Izsai bar ON							GB		*
Almásfüzitő	1757	45. pre Prepost ON			RI			RI			CB	CB
	1756	46. Prépost	RI	RI	RI	RI	RI	RI	RI	RI
		x gravel bar						GB	GB
	1755	x Dots ON							GB		*	GB
Neszmély	1751	47. Zsitva mouth	GB		GB			GB	RI		RI *	GB
	1749	48. pre Upper ON		GB	RI	GB	RI	CB	CB	CB	CB	CB
		49. Orphanage					RI	RI	RI	RI
	1748–1746	50. Upper	RI	RI	RI	RI	RI	CB	CB	CB	CB	CB
		x pre Lower bar		GB	GB	GB		GB				CB
		51. Lower	RI	RI	RI	RI	RI	RI	RI	RI	CB
		x pre Radvanyi bar	GB
		x gravel bar	GB
		52. pre Radv. ON		*						RI	RI	CB
		53. Radványi 1	RI	RI	RI	RI	RI	CB	CB	CB	CB
	1745	54. Mocsi 1	RI	RI	RI	RI	RI	CB	CB	CB	CB	CB
Süttő	1743	55. Süttői 1	RI	RI	RI	GB	RI	RI	CB	RI	RI *	CB
	1742	56. Vesszős		GB	RI	GB	RI	RI	CB	OTI	OTI *	OTI
		57. Karvai	RI	RI	RI	RI	RI	RI	RI
	1740	58. Long land ON						RI	RI	OTI	OTI *	OTI
Nyergesújfalu	1734	59. Csenke ON		RI	RI	RI	RI	RI	RI		RI	DO
	1734–1733	60. Újfalusi	RI	RI	RI	RI	RI	RI	RI	RI	RI	CB
		x Ford at Nyerges	GB		GB	GB	GB		GB	GB
Muzsla	1732–1730	x Muzslai bar	GB			GB				DO	OTI *	OTI
		61. Side-br. M. ON		RI		RI
		62. Muzslai	RI	RI	RI	DO	RI	CB	CB
		63. Small Muzsla ON		RI	GB	RI	RI	OTI	GB
Tát	1728–1722	64. Táti	RI	RI	RI	RI	RI	RI	CB	CB	CB	CB
		65. Körtvélyes	RI	RI	RI	RI	RI	RI	RI	RI	RI
	1724	66. Turán	RI		RI	RI	RI	RI	RI	RI	DO
	1723	67. Csitri		RI	RI		RI	RI	RI	RI	RI	RI
	1725–1722	68. Small Poplar ON			RI		RI	RI	GB	CB	CB	CB
		69. Poplar	RI	RI	RI	RI	RI	RI	CB	CB	CB
		70. New Ebedi ON			RI	GB	RI	RI	RI	RI	RI *	OTI
	SUMMARY
	gravel bar (GB)		6	3	5	8	1	3	7	1	0	2
	real island (RI)		13	16	20	14	21	17	12	10	8	2
	closed side-branch (CB)		0	0	0	0	1	6	11	12	13	11
	drying out side-branch (DO)		0	0	0	1	0	0	0	1	0	1
	one-time island (OTI)		0	0	0	0	0	1	0	2	3	4
	re-opened (RO)		0	0	0	0	0	0	0	0	0	0
	Year of mapping:		1771	1830–1840	183x	1857	1869–1887	1910	1934 1922–1923	1965	1980–1990	2018–2019
	Time intervals (in years):		60–70	0–10	20–25	20–30	20–30	10–15	30–40	20–25	30–40

Notes: GB—gravel bar, RI—real island, CB—island with closed side-branch, DO—island with drying out side-branch, OTI—one-time island, RO—island with re-opened side-branch. * Data from the missing section of the topographical map were replaced by maps from a sports navigation book originating from the same period, relevant page of the Danube survey was lost, but its content can be replaced from the Rauchmüller series; ¹ = These names have changed on the latest topographical nap: Radványi = Mocsi; Mocsi = Süttői; the earlier Süttői island belongs to Slovakia on this map and therefore it is not presented. Source: authors’ elaboration.

Some highlights (Table 3)

The first island along this Danube stretch is Szőnyi island (No. 42 on Figure 1). Although it was formed as directly connected to a training structure and thus, is of secondary origin, it functions as a valuable habitat today. The Harcsás-Szőnyi gravel bar was situated in the middle of the river during the 18th century and caused detrimental icy floods by stopping the ice (due to expanding in the way of a ford) in the second half of the 19th century. A law was passed to remove it in 1880 [39]; therefore, a ditch was dredged on the gravel bar, and longitudinal training structures were built in both banks to prevent the further formation of gravel bars. The southern structure was built onto the remnants of the previous gravel bar, and the dredged-out substrate was piled behind it, leading to island formation [28,38]. Four islands of various sizes compose the group along the closed side-branch today, appearing as separate islands only in cases of a high water level and dredged side-branch.

Most of the side-branches were closed due to navigation reasons since the 1880s, with only two exceptions in this section. One of the exceptions is the Prépost island (No. 45 in Figure 1). It is of typical middle-located gravel bar origin, and spindle-shaped. The island was real in the 11th century [40], its side-branch has survived the river regulations and was finally closed in 1969 in order to lead the railway tracks of the nearby alumina factory onto the island, while the upper end of the side-branch was closed by a flood protection dike, with the island acting as the source for substrate. Finally, the side-branch was converted into a fishpond.

The other exception is the Lower island of the Neszmélyi island group which also has evidence indicating its origin to have been before the 13th century [41]. Its side-branch was closed in favor of leading a road for agricultural machines to manage the croplands on the island in the 1960–1970s. This island group consists of four islands. The first Upper island is followed by the Orphanage island that has evolved in its side-branch during the second half of the 19th century. The latter is affected by a longitudinal training structure that was built during the early 1900s. It not only connects the Upper and the Orphanage islands to the bank, but even links to the Radványi and the Mocsi islands. Their size has grown due to silting up in the closed side-branch, although it was dredged during the mid-2000s.

Regulation of the Žitava river began quite early. A longish gravel bar had been formed at its one-time mouth (today Old Žitava) since the 18th century and had permanently become a small island as late as the early 20th century. Its side-branch was narrow and gradually silted up, merging the island with the bank during the first half of the 20th century.

The Danube became a state border between Hungary and Slovakia in 1920 and this resulted in a strange situation on its Neszmély–Süttő stretch. Before the consequences of the 1st World War, the Radványi and the Mocsi islands belonged to left bank settlements (Radvány—Radvaň nad Dunajom and Dunamocs—Moča). Now they officially belong to Neszmély and Süttő situated on the right bank; while the Süttői island (later gravel bar, then island again) that had belonged to Süttő village (right bank) before, became the territory of Slovakia. The story of the Süttő island is similar to the Újfalusi island, as both have evolved from middle river gravel bars, and both are small islands today, but probably were much bigger in historical times. The side-branch of the Süttő gravel bar was closed at the beginning of the 20th century and the island evolved after that time. The side-branch of the Újfalusi (later called as Nyergesi) island was closed much later, as part of managing the ford at Nyergesújfalu around the 1990s.

The island group at Tát preserves the memory of the last ford on this stretch and also has an old origin, of the 13th century [42]. The greatest among the four islands is the Körtvélyes, neighbored by the Táti and the Nyáros islands. The smallest one, the Csitri island, was already mentioned in the first paragraph of this section, as one of the two real islands along this Danube stretch. There had been a fifth island (called Turán) in the past as well, that merged with the Körtvélyes island during the second half of the 20th century, and finally, they together have practically integrated to the Táti island. Their side-branch was closed in 1949, and the Nyáras island was heightened to the mid-water level in 1950.

Key points:

• All studied islands = 29
• Real islands in 2018 = 2; (real islands in 1830s = 16)
• Islands with closed side-branch = 13 (+5 with silted up or drying out branch)
• Islands that merged to neighbor island or into the river bank or their side branch is drying out with or without regulation structures = 13
• Number of disappeared or dredged islands = 1

3.3. Esztergom–Budapest (1721–1648 rkm) River Section

The landscape of the first part of this stretch (between Esztergom and Szob-Stúrovo) was formed, besides the Danube, by the Garam (Hron) river that originates from the Nízke Tatry and transports a considerable amount of great-sized sediment. The Danube was able to transport only a part of these great-sized sediments downstream of the Garam (Hron) river mouth, while the deposition of the remaining sediment generated a series of bars until the Ipoly (Ipel) river. The Helemba island, as well as downstream temporary islands, are part of this series of bars along the middle of the river.

The most picturesque and most hydrologically diverse section of the Hungarian Danube stretch is the so-called Danube Bend, starting downstream after Szob town. The river crosses the Visegrádi mountains in a steep S-shaped valley, and creates steep curves, although not changing its character. Besides the thick gravel layer, the riverbed is formed by hard rock and marl [43], which are sometimes visible, such as in the Dömös ford (sediment swept away) or around the Helemba island (as a consequence of gravel mining).

Leaving the narrow Danube Bend, the river has built up the Szentendrei island. It was generated from the merging of several islands far before the centuries focused on in this study. The current paper does not consider its long meandering side-branch, we are focusing on the main river branch. After the two branches unite, the riverbed is stable, not susceptible to meander or change, but widening and thus, generating bars and islands.

This is the longest section in our study (73 rkm–Table 4). The islands stand alone or in pairs, not forming groups. Altogether 52 islands were compiled. The greatest number existing together at the same time is 31, during the 18th century. Today 12 real islands exist along this section.

Table 4. Changes of the islands on the Esztergom—Budapest Danube section based on mappings. Grey background highlights the real islands. Blank indicates that no island or gravel bar existed at that time.

			1771	1836–1840	1857	1880–1884	1911	1932 1922–1928	1958	1965	1980–1990	2018–2019
Settlement	fkm	Number on Figure 1 Name of the island	Bän-höltzel navig. map 1:12,240	Rauch-müller corrected 1:36,000	Pasetti 1:28,800	3rd military survey 1:25,000	Danube map 1:25,000	Water sports map (Angel) 1:25,000	Water sports map 1:40,000	Topogr. map 1:10,000	Topogr.map 1:10,000	Maxar satellite ESRI
Esztergom	1721	71. Island-top ON	RI			GB				RI	RI	RI
	1720	72. Upper				RI	RI	RI	RI
	1719	73. Ebszorító	RI	RI	RI	RI	CB	RI	RI
	1718	74. Vízivárosi	RI	RI	RI	RI	RI
Garamkövesd	1716	75. Ambói	RI	RI	GB	GB	GB	GB	GB	GB	RI **	RI
		76. exAmbó ON			RI	RI	RI	RI	RI	OTI	OTI **	OTI
	1715	77. Garam-mouth		RI	GB		GB	GB	GB	GB	GB **	GB
		x gravel bar					GB	GB	GB	GB	**	GB
	1714	78. Kövesdi	RI	RI	RI	RI	RI	RI	RI	RI	RI **	DO
		x gravel bar	GB		GB
		x gravel bar	GB		GB			GB	GB	GB	**
Helemba-Szob	1713	79. Small Helemba	GB					GB	RI	RI	RI
	1712	80. Helemba	RI	RI	RI	RI	RI	RI	RI	RI	RI	RI
	1711	x gravel bar			GB
		81. Fogarasi	RI	RI	RI	RI	RI	RI	RI	RI	DO	OTI
	1710	82. pre Dwarf ON						GB	GB	GB	RI	RI
	1710	83. Dwarf			GB			GB	RI	RI
	1709	84. Helemba bar	GB		GB			GB		RI	RI	GB
	1708	85. Szobi	RI		GB			GB	GB	GB	GB	GB
Zebegény Visegrád	1704	x gravel bar	GB		GB					GB
	1703	86. Zebegényi-			GB			GB	RI	RI	RI	RI
	1698	x gravel bar	GB	GB	GB			GB				GB
	1696	87. Visegrádi ON	RI	OTI		GB		GB	GB	GB
	1695	x gravel bar	GB	GB	GB	GB
Nagymaros	1694	x gravel bar			GB
	1693	x gravel bar	GB
		88. Yacht ON									RI	CB
		89. Long ON									RI	DO
		90. Falcon	RI	RI	RI	RI	RI	CB	DO	DO	OTI	OTI
	1692	x gravel bar			GB							GB
		x gravel bar	GB		GB			GB
Kismaros	1691–1658	91. Nobody’s	RI	RI	GB	GB	RI	RI	RI	RI	RI	RI
		92. Szentendrei	RI	RI	RI	RI	RI	RI	RI
	1691	93. Újmarosi ON								GB	RI	RI
		94. Bare-foot ON			GB			GB	GB	RI	GB	RI
	1690	95. Kismarosi		GB	RI	RI	RI	RI	CB	DO	DO	OTI
		96. Duna-Réti	RI	RI	RI	OTI	OTI	OTI	OTI	OTI	OTI	OTI
Verőce	1688	x Verőce bar	GB		GB		GB	GB
	1687	97. Verőcei	RI	GB	RI	RI	RI	RI	CB	DO	OTI	OTI
	1686	98. Kőgeszteli			GB	GB	RI	RI	CB	DO	OTI	OTI
Vác	1685	99. Martuska	RI	RI	DO	OTI	OTI	OTI	OTI	OTI	OTI	OTI
		100. Border			GB	GB	GB	OTI
	1683	101. Kompkötő	RI	RI	RI	RI	RI	RI	CB	CB	CB	RO
	1682	102. Buki	RI	RI	RI	RI	RI	CB	CB	DO	OTI	OTI
	1681	103. Gypsy bar	GB	GB	GB		GB	GB	RI	OTI	OTI	OTI
		x Groyne ON										GB
		104. Goat	RI	RI	RI	RI	RI	RI	GB	RI	RI	RI
		105. Tordai Top ON					GB	GB	GB	RI	OTI	OTI
	1680	106. Tordai					GB	RI	CB	CB	DO	OTI
	1679	107. Révész	RI	RI	RI	RI	RI	RI	CB	DO	OTI	OTI
	1678	x gravel bar	GB		GB		GB		GB	OTI	OTI	OTI
	1676	108. Égető	*	RI	RI	RI	GB	RI	RI	RI	CB	OTI
Göd	1674	109. Sződi bar ON	*		RI		GB
	1670	110. Fegyveresi	RI *	RI	DO	OTI	OTI	OTI	OTI	OTI	OTI	OTI
	1669	111. Gödi	RI *	RI	RI	RI	RI	RI	CB	CB	CB	DO
			*
			*
Dunakeszi	1666	112. Kis Szürkő ON	RI *	RI	RI		RI	GB	RI	RI	DO	RI
		113. Szürkő				RI	RI		RI	RI	RI
	1661	114. Úrréti	*	RI	GB	GB	GB	OTI	OTI	OTI	OTI	OTI
	1660	x gravel bar	*		GB
Budapest	1658	x gravel bar	*		GB		GB	CB			OTI	OTI
	1657	x Lidó	*		GB		GB	GB	RI	DO
	1655	115. Palotai	RI *	RI	RI	RI	RI	RI
	1653	116. Népsziget	RI *	RI	CB	CB	CB	CB	CB	CB	CB	CB
	1652	117. Rence	*		RI			OTI	OTI	OTI	OTI	OTI
	1654	118. Óbudai Tinies ON	RI *	RI	CB	DO	RI	RI	RI	RI	RI	RI
			RI *	RI	RI	DO
			*	RI		RI
	1652	119. Óbudai Great	RI *	RI	RI
		120. Óbudai Small	RI *	RI	CB	CB	CB	CB	CB	CB	CB	CB
		121. Spa	RI *	GB	RI
	1649	122. Margaret	RI *	RI	RI	RI	RI	RI	RI	RI	RI	RI
		123. Painter	RI *	RI	RI	RI	DO	OTI
		SUMMARY
	gravel bar (GB)		12	6	26	8	14	19	11	9	3	7
	real island (RI)		31	30	26	22	22	20	18	17	16	12
	closed side-branch (CB)		0	0	3	2	3	5	10	5	5	3
	drying out side-branch (DO)		0	0	2	2	1	0	1	7	4	3
	one-time island (OTI)		0	1	0	3	3	7	5	8	15	19
	re-opened (RO)		0	0	0	0	0	0	0	0	0	1
		Year of mapping:	1771	1836–1840	1857	1880–1884	1911	1932 1922–1928	1958	1965	1980–1990	2018–2019
		Time intervals (in years):	60–70	15–20	20–25	20–30	10–20	20–25	5–10	15–25	30–40

Notes: GB—gravel bar, RI—real island, CB—island with closed side-branch, DO—island with drying out side-branch, OTI—one-time island, RO—island with re-opened side-branch. * The Bänhöltzel navigation map [1771] shows the Danube stretch from Komárom to Vác. The Vác-Budapest part was replaced by the 1st military survey [1780–1790]; ** The missing parts of the topographical map of 1980–1990 were replaced by maps from a sports navigation book originating from the same period. Source: authors’ elaboration.

Table 4. Changes of the islands on the Esztergom—Budapest Danube section based on mappings. Grey background highlights the real islands. Blank indicates that no island or gravel bar existed at that time.

			1771	1836–1840	1857	1880–1884	1911	1932 1922–1928	1958	1965	1980–1990	2018–2019
Settlement	fkm	Number on Figure 1 Name of the island	Bän-höltzel navig. map 1:12,240	Rauch-müller corrected 1:36,000	Pasetti 1:28,800	3rd military survey 1:25,000	Danube map 1:25,000	Water sports map (Angel) 1:25,000	Water sports map 1:40,000	Topogr. map 1:10,000	Topogr.map 1:10,000	Maxar satellite ESRI
Esztergom	1721	71. Island-top ON	RI			GB				RI	RI	RI
	1720	72. Upper				RI	RI	RI	RI
	1719	73. Ebszorító	RI	RI	RI	RI	CB	RI	RI
	1718	74. Vízivárosi	RI	RI	RI	RI	RI
Garamkövesd	1716	75. Ambói	RI	RI	GB	GB	GB	GB	GB	GB	RI **	RI
		76. exAmbó ON			RI	RI	RI	RI	RI	OTI	OTI **	OTI
	1715	77. Garam-mouth		RI	GB		GB	GB	GB	GB	GB **	GB
		x gravel bar					GB	GB	GB	GB	**	GB
	1714	78. Kövesdi	RI	RI	RI	RI	RI	RI	RI	RI	RI **	DO
		x gravel bar	GB		GB
		x gravel bar	GB		GB			GB	GB	GB	**
Helemba-Szob	1713	79. Small Helemba	GB					GB	RI	RI	RI
	1712	80. Helemba	RI	RI	RI	RI	RI	RI	RI	RI	RI	RI
	1711	x gravel bar			GB
		81. Fogarasi	RI	RI	RI	RI	RI	RI	RI	RI	DO	OTI
	1710	82. pre Dwarf ON						GB	GB	GB	RI	RI
	1710	83. Dwarf			GB			GB	RI	RI
	1709	84. Helemba bar	GB		GB			GB		RI	RI	GB
	1708	85. Szobi	RI		GB			GB	GB	GB	GB	GB
Zebegény Visegrád	1704	x gravel bar	GB		GB					GB
	1703	86. Zebegényi-			GB			GB	RI	RI	RI	RI
	1698	x gravel bar	GB	GB	GB			GB				GB
	1696	87. Visegrádi ON	RI	OTI		GB		GB	GB	GB
	1695	x gravel bar	GB	GB	GB	GB
Nagymaros	1694	x gravel bar			GB
	1693	x gravel bar	GB
		88. Yacht ON									RI	CB
		89. Long ON									RI	DO
		90. Falcon	RI	RI	RI	RI	RI	CB	DO	DO	OTI	OTI
	1692	x gravel bar			GB							GB
		x gravel bar	GB		GB			GB
Kismaros	1691–1658	91. Nobody’s	RI	RI	GB	GB	RI	RI	RI	RI	RI	RI
		92. Szentendrei	RI	RI	RI	RI	RI	RI	RI
	1691	93. Újmarosi ON								GB	RI	RI
		94. Bare-foot ON			GB			GB	GB	RI	GB	RI
	1690	95. Kismarosi		GB	RI	RI	RI	RI	CB	DO	DO	OTI
		96. Duna-Réti	RI	RI	RI	OTI	OTI	OTI	OTI	OTI	OTI	OTI
Verőce	1688	x Verőce bar	GB		GB		GB	GB
	1687	97. Verőcei	RI	GB	RI	RI	RI	RI	CB	DO	OTI	OTI
	1686	98. Kőgeszteli			GB	GB	RI	RI	CB	DO	OTI	OTI
Vác	1685	99. Martuska	RI	RI	DO	OTI	OTI	OTI	OTI	OTI	OTI	OTI
		100. Border			GB	GB	GB	OTI
	1683	101. Kompkötő	RI	RI	RI	RI	RI	RI	CB	CB	CB	RO
	1682	102. Buki	RI	RI	RI	RI	RI	CB	CB	DO	OTI	OTI
	1681	103. Gypsy bar	GB	GB	GB		GB	GB	RI	OTI	OTI	OTI
		x Groyne ON										GB
		104. Goat	RI	RI	RI	RI	RI	RI	GB	RI	RI	RI
		105. Tordai Top ON					GB	GB	GB	RI	OTI	OTI
	1680	106. Tordai					GB	RI	CB	CB	DO	OTI
	1679	107. Révész	RI	RI	RI	RI	RI	RI	CB	DO	OTI	OTI
	1678	x gravel bar	GB		GB		GB		GB	OTI	OTI	OTI
	1676	108. Égető	*	RI	RI	RI	GB	RI	RI	RI	CB	OTI
Göd	1674	109. Sződi bar ON	*		RI		GB
	1670	110. Fegyveresi	RI *	RI	DO	OTI	OTI	OTI	OTI	OTI	OTI	OTI
	1669	111. Gödi	RI *	RI	RI	RI	RI	RI	CB	CB	CB	DO
			*
			*
Dunakeszi	1666	112. Kis Szürkő ON	RI *	RI	RI		RI	GB	RI	RI	DO	RI
		113. Szürkő				RI	RI		RI	RI	RI
	1661	114. Úrréti	*	RI	GB	GB	GB	OTI	OTI	OTI	OTI	OTI
	1660	x gravel bar	*		GB
Budapest	1658	x gravel bar	*		GB		GB	CB			OTI	OTI
	1657	x Lidó	*		GB		GB	GB	RI	DO
	1655	115. Palotai	RI *	RI	RI	RI	RI	RI
	1653	116. Népsziget	RI *	RI	CB	CB	CB	CB	CB	CB	CB	CB
	1652	117. Rence	*		RI			OTI	OTI	OTI	OTI	OTI
	1654	118. Óbudai Tinies ON	RI *	RI	CB	DO	RI	RI	RI	RI	RI	RI
			RI *	RI	RI	DO
			*	RI		RI
	1652	119. Óbudai Great	RI *	RI	RI
		120. Óbudai Small	RI *	RI	CB	CB	CB	CB	CB	CB	CB	CB
		121. Spa	RI *	GB	RI
	1649	122. Margaret	RI *	RI	RI	RI	RI	RI	RI	RI	RI	RI
		123. Painter	RI *	RI	RI	RI	DO	OTI
		SUMMARY
	gravel bar (GB)		12	6	26	8	14	19	11	9	3	7
	real island (RI)		31	30	26	22	22	20	18	17	16	12
	closed side-branch (CB)		0	0	3	2	3	5	10	5	5	3
	drying out side-branch (DO)		0	0	2	2	1	0	1	7	4	3
	one-time island (OTI)		0	1	0	3	3	7	5	8	15	19
	re-opened (RO)		0	0	0	0	0	0	0	0	0	1
		Year of mapping:	1771	1836–1840	1857	1880–1884	1911	1932 1922–1928	1958	1965	1980–1990	2018–2019
		Time intervals (in years):	60–70	15–20	20–25	20–30	10–20	20–25	5–10	15–25	30–40

Notes: GB—gravel bar, RI—real island, CB—island with closed side-branch, DO—island with drying out side-branch, OTI—one-time island, RO—island with re-opened side-branch. * The Bänhöltzel navigation map [1771] shows the Danube stretch from Komárom to Vác. The Vác-Budapest part was replaced by the 1st military survey [1780–1790]; ** The missing parts of the topographical map of 1980–1990 were replaced by maps from a sports navigation book originating from the same period. Source: authors’ elaboration.

Some highlights (Table 4)

The Garam (Hron) river forms a tiny delta with the Danube, giving place for the Ambó island in the 16–17th century. This island had been constantly decreasing until the 1830s, finally destroyed by the great icy flood of 1838. However, its gravel bar base has started to develop again for the 1950s, hosting 2 or 3 trees, but has become bare after about 15 years until the 1980s, when the pioneer vegetation appeared with greater coverage due to permanent low water levels. This re-born island is today a 4-hectares real island, covered with forest.

Permanently existing islands have faced significant anthropogenic effects during the past three centuries. Lonely islands that are more exposed to the river current have been used for holidays (e.g., Helemba island). Those islands that are close to cities were attached by bridges to serve recreation (Vízivárosi, Óbudai, Margaret island); while the greatest one, having areas safe from floods (Szentendrei island), hosts four settlements. All of them had originally been generated from gravel bars and obtained their current forms after the merging of one great and one small (+1–2 tiny) island. The Vízivárosi island merged with the Ebszorító island in line with silting up in the closed branch between them. The current Óbudai island consists of the one-time Great and Small islands; the branch between them was closed in 1835 to create a dockyard and port [44]. The Margaret island is a district of Budapest, formed by the Rabbit island (=Gentlemen’s or Budai island) and the Small Budai island (=Painter island).

The Dwarf and the Zebegényi islands are considered as young since they emerged around the 1920–1930s. The Bare-foot island (ON near Kismaros) became an island in the 1980s, while its neighbor, the Újmarosi island (ON) formed in the 1990s. Their generation was probably supported by the riverbed sediment wrecked during the construction of the Nagymaros hydropower plant. The construction was not finished, and the building materials were demolished soon.

The extended gravel-bar base of the Goat island is constantly shown by the maps during centuries, usually delineated as a smaller or greater island, but the changing river current and greater floods significantly decreased it during the mid-20th century. The right-bank groynes that were built in 1948–1949 opposite of Vác town [45] (1680.8 and 1681.8 rkm) generated a secondary island in the former place of the Goat island, but its side-branch is constantly narrowing, leading to it no longer being considered a real island.

There are a couple of holiday islands along this stretch, two of them being real islands today. The classic spindle-shaped Szürkő island was generated in a wide river section. Although its core is relatively old (shown already on 18th-century maps), its vegetation cover is much younger. Floods regularly cut back its development to an extent that only a bar remained at around the early 20th century, and the island has re-developed during the 1940s again. A tiny island on the same gravel bar base has developed separately next to its northern top towards the main river (called as Small Szürkő island), but they merged, although the remnants of their one-time side-branch are still visible. The island was privately owned, then nationalized in the 1950s; currently hosts holiday plots.

Downstream of the Szentendre island, after reuniting with the side-branch, the river is wide within Budapest, serving the generation of several islands. The mentioned Óbudai and Margaret islands still remain real islands today. On the contrary, the side-branch of the Népsziget (People’s island) was closed in the mid-19th century in order to build a winter harbor and dockyard [46], while the side-branch of the Palotai island silted up due to river regulation structures that were built in 1922. Two further small islands disappeared during the studied 250 years. The Spa island gave place for Roman-time remnants, but blocked navigation and was dredged, while the Rence island merged to the left bank due to silting up after creating the winter harbor. Besides them, altogether, 9 one-time islands have practically become part of the banks for now. The cross-structure of the side-branch of Kompkötő island was partly opened in 2018. The key points are as follows:

• All studied islands = 53
• Real islands in 2018 = 12 + 1 re-opened; (real islands in 1830s = 30).
• Islands with closed side-branch = 4 (+14 with silted up or drying out branch).
• Islands that merged to neighbor island or into the river bank or their side branch is drying out with or without regulation structures = 26.
• Number of disappeared or dredged islands = 10.

3.4. Summing up the Three River Sections

Altogether, 123 islands can be counted on the 150-rkm long observed Danube stretch. It is important to consider that all of these islands never existed together, but each of them was a real island during a shorter or longer period between the 18th and 20th centuries. In order to avoid misunderstandings, we compiled the results of our landscape historical investigations to present the rate of real islands compared to the other groups (i.e., islands with closed or drying out side-branch and one-time islands that became part of the riverside) at each period (Figure 2). Due to the varying aim and detailedness of each mapping, and the complexity of the changes (formation, disappearance, merging), the results only allowed us to show the trends, the directions of processes, differences among ratios, and periods of serious changes.

Altogether, 66 real islands existed at the same time along the observed 150 rkm-long Danube stretch about 200 years ago, i.e., before the large-scale water regulation works, while today this number is 18, including the two recently re-opened ones (Figure 3). This means that 73% of current potential Danube islands have closed branches or are within the stage of severe side-branch succession.

There are considerable differences in the size of the islands (Figure 4). However, due to river dynamics, the size of an island is a snapshot value; therefore, we applied the following size categories, which were measured on the latest map when the island was delineated as real: Tiny (0.1–1 ha); Small (1–5 ha); Medium (5–15 ha); Great (15–1000 ha); Giant (1000+ ha). 53 of the 123 islands are medium or great sized, which can be permanently followed during the history, we call them as “stable” islands. 40 islands were categorized as small and 29 as tiny. The tiny ones are usually established around the great islands, or behind regulation structures, and are usually short-lived as real islands.

The 53 “stable” islands list every old island that already existed before the water regulation works, but only six of them exist today as real islands, i.e., a mere 11%. Except for the youngest and smallest among them (Csitri island, 5.7 ha), all are currently used as urban (Vízivárosi, Óbudai, Margaret) or holiday islands (Helemba, Szürkő) and thus, their habitats reflect the effects of anthropogenic land uses. Even if they are partly modified with severe disturbance, they still play an irrecoverable role as ecological corridors. There are extended biologically active green surfaces on them that highly increase the urban biodiversity within the built-in areas (birds, bats, etc.). Their ecosystem services are outstanding as well. The main aim is to preserve their green surfaces, ban their further building-in, improve their habitats, and preserve their direct connection to the Danube.

Almost every island that is in a close-to-natural state and less disturbed are affected by structures and their side-arm shows different levels of silting-up, which endangers their subsistence without intervention. The side-arm of two “stable” islands (Nagy Léli, Kompkötő) have partly been re-opened recently (Figure 4).

There are currently seven real islands among the 40 small ones. All of them have established or were “re-born” during the 20th century. The “re-born” means that two current small islands were great islands in the past, but they temporarily disappeared after turning back to gravel bars by floods, and have only re-developed during the second half of the 20th century (Ambó and Kecske islands), showing the changes in the hydrodynamics. The remaining five small real islands (Zsidó, Törpe, Zebegényi, Újmarosi, Mezítlábas) developed from gravel bars in the 20th century. Their gravel-bar bases already existed in the 19th century, i.e., well before the water regulation works (Újmarosi island is questionable). They evolved to become real islands as a consequence of severe riverbed deepening and longer low water levels at dry periods. These circumstances resulted in a constant surface of bars above water level the extent of which were increased against the water flow and were stabilized by floodplain vegetation. So, the anthropogenic effects played role in their development via changing and accelerating the natural processes.

Altogether, 13 islands (Figure 4) disappeared from this 150 rkm (e.g., Lídia, Visegrádi, Fürdő, Szobi) as a consequence of dredging, or changed hydrodynamics. The Helemba bar and the Szobi island still exist in the form of gravel bars in the middle of the riverbed; therefore, they may develop into islands in the future, while the others disappeared permanently.

4. Conclusions

Smaller dikes already had been built during the Roman age, but adaptation to the natural floods had been typical for a long time. The human presence and land use along the Danube and on its islands can be detected temporarily since the Neolithic times [44]. Since the 18th century, the territorial need for agricultural land use began to increase significantly and led to the creation of the flood prevention dike system along the Danube, which resulted in a narrower floodplain. The traditional land uses of the natural habitats which remained on the flood protected area were transformed into intensive land use forms (arable lands, tree plantations), and the riverside was occupied by settlements and industries that needed a direct connection to the river. Due to these processes, the previously wide-ranged floodplain vegetation is now only concentrated on the islands and on the narrow riverbank areas. Significant river regulation interventions started in the second half of the 19th century with the purpose of (1) flood protection to allow the large volume of water pass through the river as rapidly as possible; (2) navigation route development to ensure the width and depth navigation parameters in bottleneck sections during longer periods of the year; and (3) to avoid ice floods by the removal of those natural obstacles which potentially stop the passing of the broken ice (i.e., fords, gravel bars, islands) [28,38]. Gravel bars and islands were removed by dredging, and regulation infrastructures were built to lead and keep the water in the navigation route channel part of the main flow, especially during low water-level periods. As a result of the continuously improving navigability, the waterway traffic increased. Besides several other harmful effects, navigation and urbanization contribute to the spread of invasive alien species [17], that cause massive degradation of natural habitats. Another significant problem is the decreasing low water levels caused by the intensive riverbed-erosion-induced riverbed deepening due to the regulatory structures and the consequences of the lack of sediment, which are the results of the hydropower plants as sediment traps. According to the results of the recently finished Danube Sediment project, the large and problematic [47] Gabčikovo hydropower plant and reservoir block 60% of the sediment input [48], for which sediment input has already been reduced by the chain of the upstream plants. The riverbed erosion of the Hungarian section was additionally intensified by the large volume of commercial dredging between the 1960s and 1990s. The water level deepening is well demonstrated by the measured water level curves. The riverbed formation of the Danube is the result of a long term fluvial process, the cumulative effects and the interactions of the natural processes and artificial interventions with the extreme climate conditions are causing extremely negative effects never experienced before. The significant decrease in the bedload discharge [49], together with the recent decreasing trends in the water level gradient [50], contribute to the formation of the stable and incised river channel and their consequence is a reduction in the development of gravel bars. Moreover, the adequate water depth for the ships is not ensured at a low flow regime and this called for recent interventions that severely harm the gravel bars as well as the islands.

Some historical maps that we used in our study were prepared before the river regulation work, so the presented islands are the results of natural river dynamics when their formation and elimination were nearly balanced in each mid-Danube section. Some differences in the riverbed stability are detected. The most stable section was the 2nd (Komárom-Esztergom), where medieval or even older archaeological pieces of evidence prove the existence of several islands that are mentioned in the sectional highlights. These pieces of evidence show that natural changes to the real islands happened much slower before regulation works. Many islands naturally remained real islands across hundreds of years, but there were rapid changes during the investigated period. The side branches of most of the islands have closed, silting-up processes are at different levels and many previous real islands are now one-time islands. The number of real islands was greater in the past, and this fact highlights the significantly altered river dynamics.

The development process of the younger islands established from gravel bars in the 20th century well demonstrates the joint role of natural hydrodynamics and artificial human interventions in the fluvial system. The core of their gravel bar already existed in the 19th century, well before the water regulation works. They developed to islands due to the consequences of riverbed erosion and water-level deepening, as their surface rose (almost) constantly above the water level.

Another consequence of changing natural river dynamics (less sediment, riverbed deepening, sediment excavation) is that these younger islands developed against the water flow, their size is small, and their existence as a real island is short term.

The artificially induced hydrodynamic processes are responsible for the current state of the river and its patterns. According to our observed tendencies, the dynamism of island development is significantly influenced by human interventions today. Along this river stretch, there is no chance for the development of new, bigger islands as they formed during historical times. Besides the above-mentioned tendencies, we have to face new challenges today. Climate change contributes to the rapid evolution of island dynamics, and several morphodynamic parameters changed at the beginning of the 21st century as reported by [51] in Russia. Sufficient focus on climate change adaptation with regard to issues specific to small islands is usually missing from policy issues [52]. Changing weather conditions in river basins will magnify risks emerging from extreme rainfall, temperature, runoff patterns, and altering ecological linkages [53]. Therefore, the sustainable management of the Danube should not be implemented without deep knowledge of these long-term consequences. Investigations of the above-mentioned aspects need to be taken into consideration during the planning of responsible management measures. Finding the balance among the competing demands of freshwater use, recreation, nature conservation, hydropower production, river gravel excavation, navigation and flood protection will be a real challenge, especially in the light of the current and future climate extremes. In order to conserve this unique ecological corridor and green infrastructure element in the long run, and restore its damaged alluvial forests, no further hard-tech interventions should be allowed in the fluvial system.

We conclude that the effects of any intervention into the living systems should be handled in line with its various relationships, with a thorough view of the whole system, and with a careful evaluation of their long-term consequences. The current state is the result and consequence of interventions in the past. The natural dynamics were damaged in several places, the patterns were fixed, and the riverbed was uniformized.

However, if the damages are not irreversible, then alluvial patterns and their vegetation definitely have a good regenerative potential. The plant associations are able to regenerate after stabilizing the ecological factors (e.g., water level, hydrodynamics, soil parameters) if there are propagules of the species. Such sources for regeneration might be floods that transport propagules from the vegetation of the upper river sections [54], or the seed bank, and less degraded refugees of the tributaries. Therefore, islands have great restoration potential.

The main principle is to preserve the longitudinal connectivity of the Hungarian Danube stretch because even the smallest damming would lead to irreversible consequences. In parallel, there is an urgent need for the strategic planning of a fluvial system rehabilitation program (sediment management, riverbed deepening, side-branch re-opening, etc.), with an emphasis on the hydrological changes that negatively affect the habitats of the islands and on the improvement of the riverbed dynamics. The subsistence of the naturally evolved real islands might be an indicator showing the extent of artificial interventions on a certain river stretch.

Some recommendations for the rehabilitation program are:

• Step-by-step opening of side-branches on certain river stretches, based on prioritization at system-level in the long run, with a holistic approach, considering the interests of every living being group.
• Creation of natural riverbank sections, allowing for lateral erosion.
• Incentives for starting the sediment-management planning.
• Occasional local dredging in certain closed side-branches in order to slow down their silting-up process, choosing places based on supporting monitoring data, and considering the safe handling of the potentially highly toxic material-content silt as well.