OPINION
JIM SHARP, Justice.
This is an inverse condemnation and nuisance case brought by appellees Edward A. and Norma Kerr, and over 200 other parties (the Kerrs or appellees) whose properties were damaged by flooding in the White Oak Bayou watershed.1 In four issues, appellants, Harris County Flood Control District (HCFCD) and Harris County, Texas (Harris County; collectively, appellants) contend that in its denial of their combined plea to the jurisdiction and motion for summary judgment upon its mistaken belief that the law of the case compelled denial of the plea, the trial court erred. Appellants further contend that the appellees failed to raise a question of fact with regard to each of the three elements of their takings claim, thus defeating both their takings and nuisance claims. We affirm the trial 'court’s order denying appellants’ plea to the jurisdiction.
Background
Appellees own, or formerly owned, real property in several subdivisions located in the upper White Oak , Bayou watershed. Most of these properties are homes built in the 1970s and early 1980s. Although most had never previously flooded, appellees’ properties flooded one or more times over a five-year period as a result of three severe storms: (1) Tropical Storm Frances in 1998, (2) Tropical Storm Allison in 2001, and (3) and an unnamed storm in 2002. Appellees contend that the cause of the flooding was unmitigated upstream development, coupled with appellants’ flood-con[248]*248trol measures in the White Oak Bayou watershed during the years before these storms.2
A. Brief Overview of White Oak Bayou Flood-Control Measures
White Oak Bayou is a major tributary in the Buffalo Bayou system that drains the City of Houston and the majority of Harris County, Texas. In the late 1950s and early 1960s the United States Army Corps of Engineers began planning and designing a flood-control project for the lower, more heavily populated, 10.7 miles of White Oak Bayou. The project, which entailed enlarging and lining the bayou with concrete, was not completed until the mid-1970s.
Due of severe urban flooding conditions in the remaining upstream portions of the watershed, the Corps, beginning in the early 1970s, investigated the feasibility of extending the project to the upper portion of the watershed. In 1976, the Corps produced the “Interim Report on Upper White Oak Bayou summarizing its findings as to the watershed’s recurring flooding problems and proposed a mitigation plan. This report was coordinated with the Harris County Commissioners Court (HCCC) and HCFCD.
The 1976 Corps Report confirmed the upper watershed to be prone to flooding primarily due to “the lack of adequate [bayou] capacity to carry excessive rainfall runoff away from the area without causing flooding.” Noting that the appellants had approved new development without on-site mitigation and would continue to do so, the report acknowledged these problems were further compounded by inadequate storm sewers and street drainage in the neighborhoods surrounding the upper bayou and concluded that continued upstream development would substantially increase flooding in the upper bayou watershed. Appellants concurred with the Corps’ findings and its recommendation for a major flood-control project and agreed to act as the local sponsor of such a project. Thereafter, federal funding for this project was pursued but was slow to materialize, and in November 1980, Harris County, acting through the HCCC, authorized HCFCD to implement an interim storm-water management policy that, inter alia, required all new developments in the upper bayou watershed to provide on-site storm-water detention basins. By 1988, HCFCD had completed its own “Flood Hazard Study” in which it analyzed the effects of urbanization and defined the limits of the 100-year floodplain in Harris County, Texas.
In response to the Corps’ delay in implementing its flood-control plan for upper White Oak Bayou, Harris County commissioned Pate Engineers to develop its own plan to eliminate the existing 100-year floodplain along the bayou while also providing sufficient capacity to handle addi[249]*249tional water from future development throughout the watershed. In 1984, Harris County, acting through the HCCC, formally approved the “White Oak Bayou Regional Flood Control Project” and authorized HCFCD to implement the plan (the Pate Plan). The Pate Plan was based in significant part upon HCFCD’s 1988 Flood Hazard Study, as well as FEMA floodplain maps that were also based on the 1983 study.3
The purpose of the Pate Plan was to eliminate flooding along the upper bayou, including in the vicinity of appellees’ properties, for floods up to and including a 100-year event, by expanding the bayou’s capacity to handle storm water runoff produced from existing and expected future development in the upper White Oak Bayou watershed. Because of the uncertainty of future federal funding and the current unavailability of such funding, the Pate Plan was to be funded solely through local taxes and “impact fees.” Specifically, small developments (i.e., of less than ten acres) that opted not to construct onsite detention basins were to pay $3,000 per acre “impact fees” to fund the construction of regional detention basins.
The Pate Plan was multi-phased, with the initial phase providing the necessary mitigation to prevent flooding of the developed portion of the upper bayou watershed by construction of channel improvements, including along appellees’ subdivisions, and the acquisition of regional detention sites. The plan, once fully implemented, was intended to “maintain 100-year flood protection on White Oak Bayou as future development occur[ed].”
In 1988, HCFCD informed the Corps that appellants were no longer interested in the federal flood-control project because they had developed their own. plan that would be implemented quickly with local funds.
When flooding occurred along upper White Oak Bayou in 1989, however, homeowners contacted HCFCD concerned that appellants had yet to construct any of the channel improvements called for under the Pate Plan. In response to one homeowner’s letter, HCFCD’s then-Director, James Green, acknowledged that HCFCD was “very aware of the house flooding potential in Creekside Estates South and many other subdivisions along White Oak Bayou.” Green also acknowledged that they were in the process of implementing the Pate Plan that would protect the homeowner’s subdivision from 100-year flood events and cited on-going condemnation proceedings as the reason for the delay in implementation.
According to appellants, it was during this time that HCFCD discovered that the engineering analysis used in developing the Pate Plan was either inadequate or inaccurate and, in 1990, the engineering firm Klotz Associates (Klotz), was commissioned for a new multi-phase study of existing White Oak Bayou watershed conditions. Tasked with compiling the best information, Klotz evaluated the base models and found the information in many areas to significantly differ from the FEMA floodplain maps that were based upon the 1983 Flood Hazard Study. After including new upstream development, Klotz determined that the flood flows and flood levels along the bayou were much higher than the Pate Plan showed4.
[250]*250Based upon updated models and watershed information developed during the first phase of the study, Klotz prepared an engineering report (Klotz Plan), which appellants characterize as a revised version of the Pate Plan. Klotz’s stated goal with respect to their proposed recommendations was “to find that combination of improvements which provided the most drainage value for the funds expended.”
According to appellants, they accepted the parts of the Klotz Plan that could be implemented more quickly than the original Pate Plan (i.e., constructing detention basins and earthen channel improvements to the portion of the bayou directly downstream from appellees’ homes). These proposed channel improvements extended from Cole Creek up to and past North Houstoii-Rosslyn Road, where a transition control structure designed to maintain the status quo with respect to the flood levels upstream of the structure (where appel-lees’ homes were located) would be instáll-ed.
The Pate Plan, designed to protect against 100-year flood events, called for construction of concrete-lined channel improvements for a sizable portion of upper White Oak Bayou, including near appel-lees’ subdivisions. The Klotz Plan, designed to provide protection from ten-year flood events, recommended shallower, earthen channel improvements that stopped just downstream from appellees’ properties. Thus, the portions of the Klotz Plan adopted by appellants not only protected fewer property owners but did so from less-severe flooding events. HCFCD claims that it relied upon the Klotz engineers’ certification that the revised plan complied with regulations and would not increase downstream runoff.
B. The Lawsuit
Appellees originally brought takings and nuisance claims in the 133rd District Court of Harris County against the Texas Department of Transportation (TxDOT) and several municipal utility districts, engineering firms, and real-estate developers, as well as appellants for damages arising out of the flooding of their properties during Tropical Storm Frances in 1998. See Kerr v. Tex. Dep’t of Transp., 45 S.W.3d 248, 249 (Tex.App.-Houston [1st Dist.] 2001, no pet.) (Kerr I). The trial court granted TxDOT’s plea to the jurisdiction based on sovereign immunity,5 ap-pellees took an interlocutory appeal, and this Court reversed and remanded because appellees’ pleadings sufficiently alleged the elements of a taking by TxDOT in its road design and construction. Id. at 252.
On remand, appellees settled with all of the defendants except for appellants. Appellants subsequently filed, and the trial court granted, motions for summary judgment based on sovereign immunity. The Kerrs appealed, and this Court issued an opinion reversing the summary judgment. See Kerr v. Harris Cnty., No. 01-02-00158-CV, 2003 WL 22053653 (Tex.App.[251]*251Houston [1st Dist.] 2003, no pet.) (mem. op.) (Kerr II), opinion withdrawn and superseded on rehearing, 177 S.W.3d 290 (Tex.App.-Houston [1st Dist.] 2005, no pet.). In Kerr II, appellants made some of the same arguments they make here (i.e., that they had conclusively established that they did not “intend” to take or damage appellees’ property and appellees had failed to raise a question of fact on this issue). This Court rejected that argument and held that appellees had presented sufficient evidence to raise “a clear question of fact regarding whether the actions of [appellants], in not completing the Pate plan, but in choosing to implement the Klotz plan instead, created a condition whereby the flooding of the plaintiffs’ homes was substantially certain to occur.” Kerr II, however, was subsequently withdrawn on rehearing, and was superseded by a new opinion that reversed on jurisdictional grounds because the suit had been brought in a district court instead of a county court at law. See Kerr v. Harris Cnty., 177 S.W.3d 290, 294-95 (Tex.App.Houston [1st Dist.] 2005, no pet.) {Kerr III).
After remand from Kerr II, appellees refiled their suit in the County Court at Law, where it was consolidated with two other pending takings and nuisance cases against appellants arising from flooding during Tropical Storm Allison in 2001 and an unnamed storm in 2002. Appellants filed a combined plea to the jurisdiction and motion for summary judgment arguing that the doctrine of sovereign immunity deprived the court of subject-matter jurisdiction. The trial court denied appellants’ combined plea in a written order declaring that he was compelled to do so by “Kerr v. Tex. Dep’t. Of Transp., Kerr II,” which the court believed to be “the law of the case until the Court of appeals either artfully or directly reverses itself, THUS it is binding upon this trial court.” This appeal followed.
Plea to the Jurisdiction
This Court enjoys a specific grant of appellate jurisdiction over interlocutory appeals of orders that either grant or deny pleas to the jurisdiction by a governmental unit. See Tex. Civ. Prac & Rem.Code Ann. § 51.014(a)(8) (West Supp.2012); Tex. Dep’t of Crim. Justice v. Simons, 140 S.W.3d 338, 349 (Tex.2004) (“[A]n interlocutory appeal may be taken from a refusal to dismiss for want of jurisdiction whether the jurisdictional argument is presented by plea to the jurisdiction or some other vehicle, such as a motion for summary judgment.”). A plea to the jurisdiction challenges the court’s subject-matter jurisdiction. Bland Indep. Sch. Dist. v. Blue, 34 S.W.3d 547, 554 (Tex.2000). There are two types of pleas to the jurisdiction: (1) challenges to the pleadings (i.e., whether the pleader has alleged facts that affirmatively demonstrate the trial court’s subject-matter jurisdiction) and (2) challenges to the existence of jurisdictional facts. Tex. Dep’t of Parks & Wildlife v. Miranda, 133 S.W.3d 217, 226 (Tex.2004).
When a plea to the jurisdiction challenges the pleadings, we determine whether the alleged facts as plead affirmatively demonstrate the court’s jurisdiction to hear the case. Id. Our de novo review of such challenges look to the pleaders’ intent and construe the pleadings in favor of the plaintiffs. Id. If the pleadings lack sufficient facts to affirmatively demonstrate the trial court’s jurisdiction, but do not affirmatively demonstrate incurable defects in jurisdiction, the issue is one of pleading sufficiency and the plaintiffs should have the opportunity to amend. Id. at 226-27.
[252]*252On the other hand, jurisdictional challenges accompanied by evidence are decided under a traditional summary-judgment standard in the trial court and reviewed de novo on appeal. See id. at 227-28; see also Tex.R. Civ. P. 166a(c). Thus, the burden is on the government to adduce evidence establishing that the trial court lacks jurisdiction as a matter of law. Miranda, 138 S.W.3d at 228; Porretto v. Patterson, 251 S.W.3d 701, 711 (Tex.App.-Houston [1st. Dist.2007, no pet.). Thereafter, the burden shifts to the plaintiff to demonstrate that a disputed issue of material fact exists regarding the jurisdictional issue. Miranda, 133 S.W.3d at 228; Porretto, 251 S.W.3d at 711. The defendant cannot simply deny the existence of jurisdictional facts and force the plaintiff to raise a fact issue. Miranda, 133 S.W.3d at 228; Porretto, 251 S.W.3d at 711.
Under these circumstances, we consider relevant evidence submitted by the parties when necessary to resolve the jurisdictional issues raised, as the trial court is required to do. Miranda, 133 S.W.3d at 227. Where the jurisdictional challenge implicates the merits of the case and the plea to the jurisdiction includes evidence, the trial court reviews the relevant evidence to determine if there is a fact issue. Id. If a fact question concerning jurisdiction is raised by the evidence, the trial court cannot grant the plea to the jurisdiction before that fact issue is resolved by the fact finder. Id. at 227-28. If the relevant evidence is undisputed or does not raise a fact issue on the jurisdictional question, the trial court rules on the plea to the jurisdiction as a matter of law. Id. at 228.
In reviewing a plea to the jurisdiction in which the pleading requirement has been met and evidence has been submitted in support of the plea that implicates the merits of the case, we take as true all evidence favorable to the nonmov-ant. Id. Every reasonable inference is made and any doubts are resolved in favor of the nonmovant. Id.
Discussion
In their combined plea, appellants challenged the existence of jurisdictional facts supporting all three elements of appellees’ takings claim — intent, causation, and public use. Appellees responded that the plea was contrary to this Court’s prior opinion in Kerr II, and that the plea should be denied because there were material issues of fact with respect to all three elements. As previously noted, here, the trial court denied the plea “for any and all reasons the Court of Appeals may find compelling [including,] the law of the case....”
A. “Law of the Case”
In their first issue on appeal, appellants contend that the trial court erroneously denied their plea to the jurisdiction and motion for summary judgment based upon the court’s mistaken belief that it was bound to do so by the “law of the case.”
Under the doctrine of the “law of the case,” a question of law decided on appeal to a court of last resort governs the case throughout its subsequent stages. See Hudson v. Wakefield, 711 S.W.2d 628, 630 (Tex.1986). The doctrine is based on public policy and is aimed at putting an end to litigation. Id. The doctrine does not necessarily apply, however, where the issues or facts presented in the later appeal are not substantially the same as those addressed in the earlier appeal. Id. By narrowing the issues in successive stages of the litigation, the doctrine is intended to achieve uniformity of decision as well as judicial economy and efficiency. Briscoe v. Goodmark Corp., 102 S.W.3d 714, 716 (Tex.2003). The doctrine applies [253]*253to trial courts as well as appellate courts, assuming subsequent appeal of the case. See Briscoe, 102 S.W.3d at 716.
Specifically, here the trial court stated in its order that the numerous motions pending before it, including appellants’ plea to the jurisdiction and motion for summary judgment
should be denied for any and all reasons the Court of Appeals may find compelling, and specifically in light of Kerr v. Tex. Dept. of Transp., Kerr II, which though contradictory to Aristotle’s Posterior Analytics in as much as the opinion fiats the presupposition that foreknowledge of possible future flooding is evidence of a forewill to take when a Governmental entity elects to expend its financial resources on other venues rather than proscriptively expending funds on the project at hand (a traditionally exempt exercise of legislative discretion arguably thus the robbery victim may sue for funds spent upon fire prevention and the home fire victim sue for funds spent upon police protection), IT IS, NONETHELESS, the law of the case until the Court of appeals either artfully or directly reverses itself, THUS it is binding upon this trial court.
Contrary to the trial court’s statement, neither Kerr I, which is referred to by the trial court as Kerr v. Tex. Dep’t of Transp., nor Kerr II is the law of this case. Kerr I is not binding under the law-of-the-case doctrine because that case involved different parties and different facts. See Hudson, 711 S.W.2d at 630-31 (stating that doctrine does not apply if issues or facts are not substantially same or if parties or pleadings are different). In Kerr I, another defendant who is no longer a party to this case, TxDOT, filed a plea to the jurisdiction; Harris County was not a party to that appeal, and none of the factual allegations being raised against Harris County in this case were asserted in Kerr I.
Likewise, Kerr II is not binding under the law of the case doctrine because that opinion was withdrawn and superseded by another opinion. See Scanlan v. Continental Inv. Co., 142 S.W.2d 432, 435 (Tex.Civ.App.-Galveston 1940, writ dism’d judgm’t cor.) (stating that previous rulings which were withdrawn are not law of the case); cf. Continental Cas. Co. v. Street, 364 S.W.2d 184, 188 (Tex.1963) (stating that opinions withdrawn by appellate court are no longer binding); Frizzell v. Cook, 790 S.W.2d 41, 43 (Tex.App.-San Antonio 1990, pet. denied) (stating that withdrawn opinions have no precedential value). Courts do not withdraw their opinions and issue new opinions in their place on a whim. Such actions are taken with deliberate care and concern for the jurisprudence of the State. Appellees have not offered — and we cannot conceive — of any reason why we should be bound to follow an opinion under the law-of-the-case doctrine that our learned colleagues on this Court have effectively instructed us to disregard by their decision to withdraw that opinion and substitute another in its place. Accordingly, we hold that the trial court erred when it denied appellants’ combined plea to the jurisdiction and motion for summary judgment based upon its mistaken belief that it was compelled to do so by the law of the case.
We sustain Harris County and HCFCD’s first issue.
B. Appellees’ Takings and Nuisance Claims
In their second, third, and fourth issues, appellants argue that the trial court erred in denying their combined plea and motion for summary judgment because they conclusively negated all three elements of ap-pellees’ takings claim and appellees failed [254]*254to raise a fact question with respect to any of the elements. Appellants further contend that because appellees’ nuisance claim is dependent upon their takings claim, the trial court erred in denying their combined plea and motion for summary judgment on this claim as well.
“No person’s property shall be taken, damaged, or destroyed for or applied to
public use without adequate compensation being made, unless by the consent of
such person.” Tex. Const, art. I, § 17.
Appellants are immune from any suit unless immunity is waived by statute or the constitution. See Wichita Falls State Hosp. v. Taylor, 106 S.W.3d 692, 695 (Tex.2008). Article I, section 17 of the Texas Constitution waives immunity for takings. State v. Holland, 221 S.W.3d 639, 643 (Tex.2007). Thus, if appellees do not have an actionable takings claim, appellants’ immunity remains intact.
A takings cause of action consists of three elements: (1) an intentional act by the government under its lawful authority, (2) resulting in a taking of the plaintiffs property, and (3) for public use. Holland, 221 S.W.3d at 643. The premise of Article I, section 17 is that the government should not “ ‘forc[e] some people alone to bear public burdens which, in all fairness and justice, should be borne by the public as a whole.’ ” Steele v. City of Houston, 603 S.W.2d 786, 789 (Tex.1980) (quoting Armstrong v. United States, 364 U.S. 40, 49, 80 S.Ct. 1563, 1569, 4 L.Ed.2d 1554 (1960)). Whether particular facts give rise to a “taking” of property is a question of law we review de novo. See City of Austin v. Travis Cnt’y Landfill Co., 73 S.W.3d 234, 241 (Tex.2002); Mayhew v. Town of Sunnyvale, 964 S.W.2d 922, 937 (Tex.1998).
To establish a constitutional taking, a plaintiff must show more than simply that the governmental entity’s acts themselves were intentional, City of Dallas v. Jennings, 142 S.W.3d 310, 313 (Tex.2004), or that the governmental entity was merely aware of the possibility of damage resulting from its conduct. City of San Antonio v. Pollock, 284 S.W.3d 809, 821 (Tex.2009). A plaintiff must show that the governmental entity either knows: (1) that a specific act is causing identifiable harm; or (2) that the specific property damage is substantially certain to result from an authorized government action. Jennings, 142 S.W.3d at 314. If the damage is only an accidental result of the government’s actions, there can be no intent to take the property, much less for a public use. Id. at 313-14 (discussing close link between elements of intent and public use and stating that absence of intent necessarily implies absence of public use).
Because appellees have not invoked any separate statutory or constitutional waiver of immunity with respect to their nuisance claim, they may sue appellants only “for a nuisance that rises to the level of a constitutional taking under Article I, Section 17.” Id. at 312 (holding that city retained immunity from nuisance claim because plaintiffs did not establish constitutional taking and did not assert waiver of immunity for nuisance claim). Accordingly, resolution of appellees’ takings claim is equally dispositive with respect to their nuisance claim.
1. Intent and Public Use
Relying primarily upon the affidavit of HCFCD director Michael Talbott, appellants contend that they disproved the intent element of appellees’ takings claim. Talbott’s affidavit, which explains the motivation behind and the evolution of HCFCD’s flood-control policies and deci[255]*255sions, according to appellants, clearly demonstrates HCFCD’s intent to reduce the risk of flooding in the White Oak Bayou watershed through the implementation of these policies. Talbott denies that HCFCD knew that the flooding of appel-lees’ properties was substantially certain to result from these policies. Among other things, Talbott testified that HCFCD relied on the Klotz engineers’ certifications that the Klotz Plan complied with regulations and would not increase downstream runoff. Citing to City of Keller v. Wilson, 168 S.W.3d 802, 829 (Tex.2005), appellants argue that their reliance upon the Klotz engineers’ certifications negates any suggestion of intent.
Appellees counter that Talbott’s affidavit (1) only applies to HCFCD, not Harris County,6 and (2) is insufficient to conclusively establish lack of intent because the affidavit contains baseless, conclusory opinions that amount to nothing more than a sworn denial of appellees’ legal claims. See Pollock, 284 S.W.3d at 816; Burrow v. Arce, 997 S.W.2d 229, 235 (Tex.1999). Appellees further contend that City of Keller is factually distinguishable and appellants’ reliance on this case is misplaced.
Alternatively, appellees argue that even if appellants met their burden by adducing evidence negating the element of intent, the trial court’s denial of their plea was proper because appellees brought forth sufficient evidence to raise a question of fact with respect to this issue. In particular, appellees argue that there is a question of fact as to whether appellants knew that their approval of upstream development without sufficient mitigation, coupled with their adoption of the Klotz Plan, was substantially certain to lead to the flooding of appellees’ properties. Appellees cite to the engineering studies that were provided to, or prepared for appellants (e.g., the 1976 Corps Report, the 1984 Pate study, and the 1990s Klotz studies), statements by county officials, and the testimony of appellees’ expert as evidence of appellants’ intent.7
Appellants argue that, although a required element of a constitutional taking, public use is merely a tag-along factor closely linked with the element of intent. Appellants further contend they have conclusively negated the element of public use for the same reasons they conclusively negated the element of intent. Jennings, 142 S.W.3d at 313-14.
We conclude that appellees presented evidence sufficient to raise a question of fact with respect to the elements of intent and public use.8 Appellees present[256]*256ed numerous of the Corps’ studies and reports, as well as those by engineers hired by appellants that, appellees contend, demonstrate that appellants knew that their role vis a vis the watershed (e.g., approval of upstream development without sufficient mitigation) was substantially certain to result in the flooding of appellees’ properties. Appellees’ flood expert, Dr. Larry Mays, reached the same conclusions in his 2009 expert report.
Citing to the 1976 Corps Report, appel-lees argue that appellants were aware that their approval of upstream development without sufficient mitigation was substantially certain to flood appellees’ properties as early as 1976. This Report informed appellants that the flooding in the upper White Oak Bayou watershed was due primarily to “the lack of adequate [bayou] capacity to carry excessive rainfall runoff away from the area without causing flooding,” and that this caused the bayou to overflow its banks and flood adjacent properties. The report further stated these “flooding problems [were] compounded by continuing [upstream] urbanization which increases and accelerates the runoff from rainfall ... and severe localized flooding caused by inadequate storm sewers and street drainage.”
HCFCD’s 1988 Flood Hazard Study confirmed that the floodplain along upper White Oak Bayou had increased significantly since the 1976 FEMA map and the 1976 Corps Report. The study’s findings were consistent with the Corps’ predictions that upstream development would necessarily result in an expanded floodplain and increased flooding along upper White Oak Bayou.
Appellants adopted the Pate Plan the following year. Appellees argue that the
fact that the appellants adopted the Pate Plan to address these flooding problems is evidence that they knew that upstream urbanization caused increased runoff and accelerated flooding downstream, and if they did not take action to provide mitigation that could accommodate past and future upstream development, it was substantially certain that downstream properties (including appellees) would flood.
Dr. Mays opined that despite the fact appellants knew that its approval of upstream development without sufficient mitigation was substantially certain to flood appellees’ properties as early as 1976 (a fact which was reinforced in 1983), “numerous developments in the upper White Oak Bayou watershed were approved after 1976 by the County without detention,” citing to appellants’ expert, Melvin Spinks. After reviewing the Pate Plan, which also stated that increased runoff from unmitigated development would exceed the runoff capacity of the bayou, Dr. Mays concluded that appellants knew in 1984 that approving unmitigated development and not implementing the mitigation measures of the Pate Plan was substantially certain to result in the flooding of appellees’ homes.
Nevertheless, appellants subsequently chose to implement a scaled-back version of the Pate Plan — the Klotz Plan — which undeniably provided less protection to fewer property owners. Citing to City of Keller, the County argues that it relied on the certification of the Klotz engineers that implementing this revised, scaled-back version of the Pate Plan, would not increase downstream runoff, and thus, could not have manifested the requisite intent for a [257]*257takings claim. City of Keller, 168 S.W.3d at 829.
City of Keller, however, is factually distinguishable. In City of Keller, three sets of engineers certified that the enacted flood-control plans would not increase downstream flooding. 168 S.W.3d at 829. Although the Klotz engineers also certified that their plan would not increase downstream flooding, unlike the plaintiffs’ property in City of Keller, appellees’ properties are located upstream from the channel improvements recommend by Klotz. Thus, the certifications relied upon by the County in this case are irrelevant with respect to whether they knew that increased flooding on appellees’ property was substantially certain to occur if they implemented the Klotz Plan. Moreover, in City of Keller, the plaintiffs’ expert testified that flooding was inevitable, but offered no testimony that the city knew that flooding was inevitable. Id. Thus, the plaintiffs merely proved that the city might know, not that it did know that “the plans it approved were substantially certain to increase flooding on the [plaintiffs’] properties.” Id. at 830. Here, appellees’ expert, Dr. Mays, opined that flooding was substantially certain to occur and that Harris County knew that it was, based, in part, upon the Pate Plan, the 1976 Corps Report, as well as other studies.
The dissenting opinion dismisses Dr. Mays’s opinions on this issue as concluso-ry — based in part upon deposition testimony Dr. Mays gave in this case in 2001. Notably, by the time the plea to the jurisdiction and motion for summary judgment were decided eight years after his deposition was taken, the Doctor’s opinion on the matter had evolved to incorporate these pleadings into his analysis. As the dissent acknowledges, Dr. Mays’s June 2009 report supplemented his January 2001 report and amended his 2007 supplemental report. Contrary to the dissent’s suggestion, Dr. Mays did not “assume” that on-site detention facilities in the watershed were inadequate “merely because flooding subsequently occurred,” as appellants argue. Dr. Mays opined that the upstream development failed to include adequate detention based upon information included in the report of Melvin Spinks, one of appellants’ experts.
Appellees also argue, inter alia, that Talbott’s own affidavit creates a question of fact with respect to HCFCD’s knowledge and reveals why the issue of intent is generally one of fact that cannot be decided as a matter of law in the present case. In particular, appellees point to the portion of the affidavit in which Talbott opines that, “what [HCFCD] ‘knows,’ or whatever anyone ‘knows,’ in the context of planning and implementing complex flood risk reductions projects is a very qualified concept.” Talbott acknowledges that HCFCD’s knowledge has “changed significantly over time,” and that “what [HCFCD] ‘knew’ about this watershed in 1984 and what [HCFCD] knew in the early 1990s or even today ... is a function of expanding information and technical capabilities.” Appellees argue that construing Talbott’s statements about HCFCD’s ever-changing knowledge in appellees’ favor— as required by our standard of review— reveals that these statements raise a fact issue about what HCFCD knew about the potential for flooding and when HCFCD knew it.
Appellants respond that appellees miss the point of Talbott’s statement, which is to explain that perfect knowledge does not exist regarding the behavior of storm-water runoff in a watershed and an engineers’ understanding of any particular watershed is continually changing based upon new technology and observations of actual events. According to appellants, this high[258]*258lights the “difficulty in predicting the effects of any one factor and illustrates the risk of attributing intent and fault with hindsight.” Be that as it may, Talbott’s statements clearly highlight the factual complexity surrounding this issue and the difficulty of pinpointing precisely what appellants knew about flooding in the upper watershed and when they knew it.
While none of this evidence standing alone may be sufficient to raise a question of fact, when considered together and taken as true, and after resolving every reasonable inference in appellees’ favor, ap-pellees’ evidence is sufficient to raise a question of fact with respect to appellants’ intent to take appellees’ property for a public use. See Miranda, 133 S.W.3d at 227. This is particularly true in the present case, where we have conflicting expert testimony as to what appellants actually knew and when they knew it, and, unlike in City of Keller, there are no expert certifications for the appellants to reply upon to negate intent.9
We hold that appellees raised a question of fact with respect to the intent and public use elements of their takings claim.
2. Causation
Proximate cause is an essential element of a takings case. “[W]ithout causation, there is no ‘taking.’ ” Tarrant Reg’l Water Dist. v. Gragg, 43 S.W.3d 609, 615 (Tex.App.-Waco 2001), aff'd, 151 S.W.3d 546 (Tex.2004).
Appellants contend that they disproved the causation element of appellees’ takings claim, relying primarily upon the affidavits of three of their experts — Talbott, Spinks, and Andrew Yung — all of whom addressed the issue of causation. According to appellants, this evidence demonstrates that the flooding of appellees’ properties was attributable to the severity of the rainfall associated with the three storm events at issue in this case, as well as the inadequate storm sewers and other local drainage problems.
Appellees’ evidence on this issue consists primarily of the reports prepared by their expert, Dr. Mays, and the materials upon which those reports are based. In his June 2009 report, which supplemented his January 2001 report, and amended his 2007 supplemental report, Mays explicitly eliminated other possible causes of the flooding, including appellants’ experts’ claims that the flooding was possibly the result of local drainage systems that were unable to handle the unusually high amounts of rainfall associated with these storms. Dr. Mays also reaffirmed in his supplemental report that appellants’ approval of unmitigated land development, as previously discussed, caused the flooding of appellees’ homes in all three storm events. Notably, the affidavits of Talbott, Spinks, and Yung neither address Mays’s [259]*2592009 report nor do they opine that the unmitigated development demonstrated by Mays was not a cause of the flooding.
After resolving every reasonable inference in appellees’ favor, we hold that appellees presented evidence sufficient to raise a question of fact with respect to this element as well. Having held that appel-lees raised a question of fact with respect to all three elements of their takings claim, we further hold that the trial court did not err in denying appellants’ plea to the jurisdiction with respect to both their takings and nuisance claims.
Conclusion
We affirm the trial court’s order denying appellants’ plea to the jurisdiction.
Justice BROWN, concurring and dissenting.